The Seko Languages of South Sulawesi: a Reconstruction

advertisement
The Seko Languages of South Sulawesi: a Reconstruction*
Tom Laskowske
SIL International
tom_laskowske@sil.org
Prior to this article, the only historical reconstruction to include a Seko language was Mills’
1975 reconstruction of Proto South Sulawesi (PSS). However, his failure to notice contrastive
word-final vowel length led him to erroneously conclude that no South Sulawesi language exhibits
a reflex of Proto Austronesian (PAn) *q. Hence *q was omitted from his PSS reconstruction.
Countering that conclusion, Sirk (1989) pointed out evidence for the existence of PSS *q. Seko
languages are therefore a test for PSS *q.
Since Mills’ reconstruction, Sirk addressed the question of what characteristics define PSS.
He and other researchers noted similarities between the Badaic languages of Central Sulawesi and
South Sulawesi languages, particularly Seko,1 but questions about genetic relationships persisted,
which motivated me to undertake this reconstruction.
This paper seeks to establish a low-level node called Proto Seko, which should provide a
basis for eventually clarifying Seko’s position in relation to other Sulawesi languages. I herein
present a reconstruction of Proto Seko phonology on the basis of data from four closely related
languages in northern South Sulawesi (Seko Padang, Seko Tengah, Panasuan, and Budongbudong), based on several years of fieldwork in this area. Besides retaining a reflex of PAn *q,
Seko languages share a quickness feature which corresponds to PAn *e.
1.
Introduction
The four languages that belong to the Seko sub-group are, or until recently were, located
along the upper reaches of the Karama River, the third largest river system in Sulawesi, a major
island in north central Indonesia. The river flows west, emptying into the Strait of Makassar
north of Mamuju (map - Figure 1).
These languages, according to data from the 1980s, number 5000 speakers of Seko Padang,
2500 of Seko Tengah, 1500 of Panasuan, and 11 households of Budong-budong speakers (with
only 5 households where both adults are Budong-budong speakers). According to Panasuan oral
history, Budong-budong is a recent split from Panasuan.
After a brief overview of the common synchronic phonology of Seko languages in Section
(§) 2, I will discuss changes that have taken place between Proto Malayo-Polynesian (PMP) and
Proto Seko (PSek) in consonants (§3) and vowels (§4), with special attention to PMP *e
(hereafter symbolized *ə). Next is a brief look at the present-day reflexes in Seko languages (§5),
and hints toward classification of PSek are presented in §6. Reconstructions are listed in §7.
2.
Seko Synchronic Phonology
2.1
Overview of basic features
The common basic elements of the phonology of Seko languages are stated in the following
list.
*
I would like to acknowledge David Mead for reading drafts of this paper. I am grateful for his many helpful
comments, and for his guiding and encouraging me in my initiation into Austronesian historical linguistics. Thanks
also to Michael Martens for input relating to Badaic languages, and for keeping me pointed in the right direction.
1
Sirk (1989:55-57), Martens (1989:22-23, 31-33), Mead (2003:130-3, 136-8)
2
Figure 1. Location of Seko languages
SULAWESI
CENTRAL SULAWESI
UMA
BADA'
Lake
Poso
WEST SULAWESI
RAMPI
Budong-budong
Hono'
SEKO PADANG
Strait of
Makassar
Panasuan
Karama R.
Talondo'
SEKO
TENGAH
Lodang
KALUMPANG
SOUTH SULAWESI
Mamuju
Rantepao
1. Five vowels:
i, e, a, o, u
2. Fourteen consonants:
- vl stops:
- vd stops:
- nasals:
- fricatives:
- liquids
- semivowel
p, t, k, ʔ
b, d
m, n, ŋ
s, h
l, r
y
Bay of
Bone
Recent loan phonemes are: /g/, /w/, and palatal affricates /c/, and /j/.
In Panasuan, /c/ is an allophone of /t/ preceding high vowels. In the Hono' dialect of
Seko Padang, retroflexed lateral flap [ ɭ ] is an allophone of /l/, but unmarked in the
orthography.
The phoneme /y/ does not occur word initially, except in loans.
3
3. Two suprasegmental/prosodic features:
- vowel length:
i:, e:, a:, o:, u:
- vowel quickness:
ì, è, à, ò, ù
4. Words may end with a vowel, a long vowel, glottal, /k/, or /ŋ/. Long vowels only occur
word finally. Glottal only occurs in root final position contrastively. It also occurs
predictably between identical vowels, but is not symbolized in that environment.
5. Sequences of two vowels are common, each vowel filling a syllable slot.
6. Only one consonant cluster pattern is found root medially, namely -NC-. The pattern
manifests itself variously among the languages, but geminate consonants are a common
manifestation.
7. Stress is penultimate.
8. Closed syllables are common. Closed syllable structure is maintained when adding stressshifting suffixes.
9. Syllable structure is fixed/assigned in the lexicon at the morpheme level. Stressed
syllables contain a vowel nucleus followed by a coda. The coda is either a consonant,
resulting in a closed syllable, or an empty timing mora, resulting in an open syllable. We
could call the empty timing mora a DUMMY CODA or a CODA BLOCKER. Unstressed
syllables may have a coda, but it is not required.
10. Possessive suffixes, which all begin with consonants and are stress-shifting, only attach
to stem-final codas. Failing to find one, a homorganic nasal ligature is inserted between
the stem and the possessive suffix.
2.2
More about prosodic features
A further word of description is in order about some of the more exotic features of Seko
phonology. We will start with the prosodic features of vowel length and vowel quickness.
VOWEL LENGTH occurs in word final position, where it is contrastive. Length can occur on
any of the five vowels. Word stress does not shift, but remains penultimate. Consider the
following two-syllable roots from Seko Padang.
(1) a.
a'.
mata
mata:
‘eye’
‘raw’
(2) a.
a'.
hulu
hulu:
‘body hair, feathers’
‘shin’
(3) a.
a'.
piso
siso:
‘machete’
‘snail’
4
(4) a.
a'.
lose
sise:
‘lazy’
‘tooth’
(5) a.
a'.
api
rapi:
‘fire’
‘twin’
VOWEL QUICKNESS is contrastive only in the stressed syllable (penultimate position). As a
point of reference, we could say that all vowels in closed syllables are QUICK. Normally, a
penultimate closed syllable nucleus is followed by two consonants, like the /a/ is followed by
/nd/ in example (6) from Seko Tengah. In such words it is not necessary to mark the /a/ as quick.
The coda of the first syllable is the first consonant, /n/, and the onset of the second syllable is the
second consonant, /d/.
(6)
tanduk
‘horn’
What is unusual about Seko languages on this point is that they can have closed penultimate
syllables that are followed by only one consonant. In example (7), the first syllable man- (written
màn-) is closed, ending with /n/, and the final syllable -na' begins with /n/. The /n/ thus straddles
two syllables, functioning as both coda of the first syllable and onset of the second. We refer to
such consonants as AMBISYLLABIC. This stands in contrast to (8), whose penultimate syllable is
open, and also contrasts with (9), whose lengthened medial consonant is analyzed as a cluster.
These examples are from Seko Padang.
(7)
(8)
(9)
màna'
mana'
mannang
‘quiet’
‘inheritance’
‘remain’
Besides having shorter timing than vowels in open syllables, quick vowels are reduced
(centralized) in quality, most noticeable for /à/, pronounced schwa [ə]. Front vowels /ì/ and /è/
are somewhat reduced to [ɩ] and [ɛ] respectively, whereas reduction in vowel quality is barely
noticeable for /ù/ and /ò/. For further description of quick vowels and ambisyllabic consonants in
Seko Padang, refer to Laskowske (1993).
2.3
More about maintaining closed syllables
Maintaining closed syllable structure is an important feature of Seko phonology. This is
most readily demonstrable when suffixes are added to stems ending in a consonant. For example,
the intransitive verb stem in (10a) ends in a /k/. The non-stress-shifting echo vowel -i in Seko
Padang is lost when nominalizing suffix -ang is added (10a'), and /m/ in the prefix automatically
goes to /p/. But the main thing to note is that, while stress shifts from syllable ku in (10a) to
syllable tik in (10a') (stressed syllables underlined), no resyllabification takes place. tik remains a
closed syllable, in which /i/ is a quick vowel, and /k/ is ambisyllabic, so we mark it in the
orthography with a grave diacritic over the vowel.
(10) a.
a'.
ming-kutik-i
ping-kutìk-ang
‘ask’
‘question’
5
2.4
More about codas and possessive suffixes
The point of talking about codas in syllable structure is that it helps us explain two
phenomena in Seko phonology. The first is when discussing what causes contrastive quick
vowels. Quick vowels within roots cannot be predicted based on any sequence of segmental
phonemes, so the feature must have already been assigned in the lexicon. If such assignment is
related to syllable structure, a corresponding structure should be found in ancestral or related
languages. Such a rule might read:
A consonant between two vowels is tagged plus or minus coda, where the first vowel Vα is
the nucleus of a stressed syllable, symbolized:
C > C+/-coda / Vα_V
The second phenomenon related to codas in syllable structure is the insertion of a
homorganic nasal ligature preceding stress-shifting suffixes that begin with a consonant (all of
which happens to be the entire set of possessive suffixes), and following stems that end in vowel
/i/, /a/, or /u/. The rule might read:
The onset of a consonant-initial suffix becomes a nasal-consonant cluster when following a
vowel, where the nasal takes on the point of articulation (α) of the following consonant, and where
the vowel is one of the historic vowels, namely /i/, /a/, or /u/, symbolized:
-Cα > -NαCα / V_
It would be most elegant if just one ‘rule’ proved adequate to account for both quick vowels
and nasal insertion. This is what we will attempt to do in the following paragraphs.
By focusing on the concept of coda, we go to the heart of the difference between open and
closed syllable structure. The simple solution of assigning a +/- closed syllable feature from the
lexicon would be adequate for vowels preceding root medial single consonants, but not for stem
final vowels. We need some way to distinguish the behavior of vowels /i/, /a/, and /u/, which
trigger a nasal ligature (underlined) on possessives, from vowels /e/ and /o/ and long vowels,
which act like consonants in not triggering a nasal ligature.
uhai -na
tai:
-na
lame -ta
sise: -mu
mata -ku
lila: -ku
piso -ki
lindo: -ku
hulu -na
hulu: -na
sadang -ku
koko' -mu
anak -ku
> uhainna
> taina
> lameta
> sisemu
> matangku
> lilaku
> pisoki
> lindoku
> hulunna
> huluna
> sadangku
> koko'mu
> anakku
‘his drink’
‘its intestines’
‘our (INCL) potatoes’
‘your teeth’
‘my eye’
‘my tongue’
‘our (EXCL) machete’
‘my forehead’
‘its feathers’
‘her shin’
‘my jaw’
‘your leg’
‘my child’ SekT
6
We could say that, because long vowels are derived historically from final *q, and /e/ and /o/
are derived from final *ay and *aw, the possessive nasal rule is consistent from a historical
perspective, in that nasals are inserted after former vowels but not after former consonants. But
that doesn’t explain why synchronically /e/ and /o/ and long vowels act like the other vowels
when shifted over to the stressed syllable due to suffixation.
I propose that coda slots can be filled by more than just consonants. In the case of word final
vowel length, the coda slot is filled by a timing mora (L) that takes on the segmental features of
the preceding vowel. In the case of root final /e/ and /o/, the coda slot is filled by a dummy coda
(D). When suffixes are added to words ending in L or D, these suprasegmental coda features
effectively block consonants from filling the coda slots which they occupy.
3.
Consonants, from PMP to Proto Seko
3.1
Introduction
The following general comments can be made about the consonant reflexes in PSek:
1. Most final consonants have merged or become lost. They will therefore be treated
separately in §3.14.
2. Any reflexes of PMP *c are subsumed under *s. They are so infrequent, I think I have no
correspondences to reconstruct.
3. The first consonant in non-nasal clusters has been lost, while the reflex of the second
consonant sometimes becomes ambisyllabic. In such a case, the preceding vowel will be
marked with a grave accent, indicating quickness.
Table 1. Proto Seko reflexes of PMP consonant phonemes in initial and medial positions
PMP
PSek
*p
*p
PMP
PSek
*m
*m
*t
*t
*k
*k, ∅
*n, ñ
*n
*q
*∅ , :
*ŋ
*ŋ
*b
*b, h
*N
*n
*d
*r, d
*l
*l
*r
*∅
*R
*∅
*s
*s
*w
*h, u
*Z
*d, r
*j
*r
*nZ
*(n)d
*nj
*h
*(n)d *∅
*g
*k
*y
*y
Before dealing with the phonemes that merit discussion (§3.2 – 3.12), I will list the
straightforward cases, with illustrations. PMP forms cited are taken from Blust (1999 or 1995)
where possible, and those are usually unmarked in the examples. I have tried to explicitly
reference other sources.2
2
Refer to the associated source for the following abbreviations PWMP (Proto Western Malayo-Polynesian), Blust
1995. PPh (Proto Philippine). PIn (Proto Indonesian) or PSS (Proto South Sulawesi), Mills 1975. PPUS (Proto Pitu
Ulunna Salu) or PMAT (Proto Mambi-Aralle-Tabulahan), Smith 1993. PKP (Proto Kaili-Pamona) Martens 1990.
PMun (Proto Muna) van den Berg 1991.
7
(11)
(12)
(13)
(14)
(15)
(16)
(17)
PMP
*p *panij
*hapuy
*t *tasik
*kutu
*s *susu
*asu
*m *mata
*lima
*ñ *ñawa Dahl
*n *naquR PPh3
*tunu
*ŋ *ŋaŋa4
*buŋa, buŋaŋ
(19)
*li(n)juq
*ala
*h *hasaq
*duha
*y *bayad Zorc '95
3.2
PMP *k
(18)
*l
‘wing’
‘fire’
‘sea’
‘head louse’
‘breast’
‘dog’
‘eye’
‘five’
‘breath’
‘descend’
‘burn’
‘agape’
‘flower’
‘forehead’
‘get, fetch’
‘sharpen’
‘two’
‘pay’
PSek
*paniʔ
*api
*tasik
*kutu
*susu
*asu
*mata
*lima
*mi-na-naha
*nau:
*maN-tunu
*mi-ŋaŋa
*buŋa
buŋa-buŋa
*hùŋaŋ
*lindo:
*maN-ala
*asa:
*i-dua
*maN-bayaʔ
‘breathe’
‘go downstream’
‘roast’
‘agape’
‘decorative shrub (sp)’
‘decorative flower’
‘primordial fruit
or its flower’
‘file teeth’
PMP *k is generally reflected in PSek as *k, but in a significant number of words it is lost. I
have not detected any general rule that conditions the loss, except that root initially only two or
three forms reflect loss of PMP initial *k-, versus approximately 90 forms that retained it,
whereas root medially 11 out of 41 (27%) lost *-k-. After a couple of examples where *k is
retained (20), I list all known examples of PSek words where *k is lost (21).
(20)
(21)
3
PMP
*kulit
*bikaq
*kita
*kua PSS
*ka
*lakaw
*dalikan
*sakit
*kaka
*daki
PSek
‘skin’
*kuliʔ
‘split’
*maN-bika:
‘see’
*maN-ita
‘say’
*maN-oa
‘if’
*ia5
‘go’
*lao
‘three-stone fireplace’ *laliaŋ
‘sick, painful’
*sai ‘illness’
‘elder sibling’
*kaa
‘fine dirt (on skin)’ *rai SekP
*t > *∅ irregular
Charles (1973)
Dempwolff (1938) and Blust pers. com. to David Mead
5
Without external evidence, it's hard to say it this point what happened. It is possible *k was not lost in initial
position, but rather PMP *ka > *i-ka > PSek *ia, in which case it would have been lost in medial position.
4
8
?*dikaw
*ikəj PWMP
*dakut
*buku
*siku
*nakaw dbl
‘you’
‘cough’
‘grab’
‘node, joint, knuckle’
‘elbow’
‘steal’
*dio
from where Mal dikau?
*m-eiʔ
*maN-rauʔ
*buu:
‘bone, core’
*siku
[SekP, SekT siu]
*mu-ma-nau
PSek is not, however, the only proto language to have lost *k. Take, for example, the PSS
word *kande (cognates also found elsewhere in Sulawesi), which in PSek has two reflexes,
*kande: and *ande:. Data presented by Mills (1975:723) shows that reflexes with and without /k/
are found throughout South and Central Sulawesi. Example (22) lists derivations found in Seko
Padang. Only the last three items have the /k/ reflex.
(22)
3.3
m-anne:
maN-anne:
maN-anne-i:
maN-pa-anne:
maN-pu-m-anne:
paN-anne:
anne-ang
ka-m-anne-ang
pa-anne:
pi-anne:
k-in-anne:
maN-pa-kanne:
maN-paN-pa-kanne:-ing
‘eat (intransitive)’
‘eat (transitive)’
‘eat from’
‘feed to’
mampaanne: saha asungku
‘have appetite for’
‘side dish’ (Mal lauk pauk)
‘plate, food’
‘board (as in, “room and board”)’
‘able to cut’
‘eater, (house on) fire’
‘cooked rice, food’ (-in- infix)
‘feed’
‘feed with’
PMP *q
PMP *q is lost root initially in PSek (23), and root medially (24). In some words, *-qmetathesized with the final vowel which then became lengthened in PSek (25). For a description
of word final *q, see §3.14.
(23)
(24)
(25)
PMP
*qanitu
*qahəlu
*quban
*laqia
*bituqən
*paqit
*buqaya
*tuqəD
*ma-tuqah
*taqi
*paqa
‘spirit, soul’
‘rice pestle’
‘gray hair’
‘ginger’
‘star’
‘bitter’
‘crocodile’
‘tree stump, stubble’
‘mature, old’
‘feces’
‘thigh’
PSek
*anitu SekP
*alu
*ubaŋ SekP
*laia
*bintoeŋ
*mi-paiʔ
*buaya
*toaʔ
*ma-tua:
*tai:
*paa:
‘spirit, ghost, goblin’
‘base of tree’
‘mature, old, hard’
also, ‘guts’
[but, paa archaic, paanna]
9
3.4
PMP *b
PMP *b is reflected in PSek as *b and *h in both initial and medial positions. No variation is
found between /b/ and /h/ reflexes in Seko languages, so all PMP *b that became PSek *h,
presumably through transitional *w, happened pre-PSek.
Two generalizations can be made. The first is that PMP *bVb- sometimes became PSek
*bVh- and sometimes *hVh-, but never *hVb-. So in (28) we find PSek reflexes *bahi and
*baha, but never anything of the pattern **habi or **haba.
The second generalization is that many PSek words with intervocalic *b can trace back to
something that blocked the shift to *h, for example a consonant cluster (29b-c). Other
intervocalic /b/ words in Seko languages follow a frozen PMP prefix (29d). Finally, some words
had sound changes happen after PMP *b went to *h intervocalically, and missed the shift, for
example, metathesis (29e), and loans, like (29f).
In conclusion for PMP *b, besides the intervocalic environment in general, no other
conditioning factor is evident for the shift toward Proto Seko *h. In many cases, the shift to *h,
or lack thereof (29g, 29h), remains unexplained.
(26)
(27)
(28)
(29)
a.
b.
c.
d.
e.
f.
g.
h.
PMP
*təbuh
*bibiR
*tabuni
*babaw
*bubu
‘sugar cane’
‘lip’
‘afterbirth, placenta’
‘above’
‘fish trap’
PSek *h word medial
*tùhu
*hihi
‘edge, perimeter’
*tahuni SekP
*i-aho
*huhu
PMP
*batu
*bulan
*bulu
*buluq
‘stone’
‘moon’
‘body hair, feathers’
‘slender bamboo’
PSek *h word initial
*hatu
*hulaŋ
*hulu
*hulo:
PMP
*babuy
*baba
*bənaŋ PWMP
‘pig’
‘carry’
‘thread’
PSek *b word initial
*bahi
*maN-baha
*bànaŋ
PMP
*baqbaq
*ti(m)baR PIn
*bəlbəl
*baŋbaŋ
?*bu(ŋ)kuj Mills
*tabaroʔ PKP
*sa-Ribu Dahl
*təbək
‘mouth’
‘throw’
‘plugged, stupid’
‘butterfly’
‘back’
‘sago palm’
‘thousand’
‘pierce, stab’
PSek *b word medial
*baba:
‘door’
*maN-tibe
‘throw away’
*bebe SekP
‘mute, stupid’
*kalibambaŋ
u > e unexplained
*keboʔ
X
tabaro:
expected r > ∅
*sàbu
*maN-tòbok
‘poke’
tobok
‘kris’ [borrowed]
maN-tobok
‘stab’ [
" ]
10
3.5
PMP *d
PMP *d went to *r in PSek, but some words apparently never made the shift. Perhaps the
migration/shift was incomplete at that point. The migration was more complete word medially
than word initially. Example (30) shows root initial *d > *r, and (31) word medial.
(30)
(31)
PMP
*dasan PIn '81 ‘house’
*dakit
‘raft’
*dapuR
‘hearth’
PSek
*rasaŋ
*rakiʔ
*rapu SekP
X
dapoʔ SekP
*tuduR
*kudən
*adəp PWMP
*ŋuda dbl
*ma-turu
*koriŋ
*aruʔ
*ma-ŋura
‘sleep’
‘cooking pot’
‘chest’
‘young’
‘clan’, (‘hearth’ - Lodang’)
‘hearth’ (later loan)
‘lie down’
The most common exception is where PMP *d remained *d in PSek (32, 33). I know of no
explanation for root initial *d to remain, except that it is more resistant to change than in the
medial position, the same phenomenon we observed for PMP *b (§3.4). Root medially, however,
the *d can be explained. The shift was probably blocked when following a consonant. In (33a)
the pattern is clear. Probably (33b) and (33c) were inherited from the forms that had a consonant
cluster, where the *n consequently got dropped after the *d-to-*r shift took place.
PMP
*duha
*dahun
*duluR
*duRi
*datuʔ Zorc '95
(32)
(33)
a.
b.
c.
‘two’
‘leaf’
‘accompany’
‘thorn’
‘chief’
*tanduk
‘horn’
*pa(n)daŋ Mills ‘meadow’
*sə(n)du PAn
‘hiccup’
PSek
*i-dua
*dauŋ
*maN-dulu
*dui
*datu:
[note: karo'a: ‘eight’]
‘work together, help’
‘king’
*tanduk
*padaŋ
*sùdu SekP
Another exception is where PMP *d went to *l in PSek (34a). The shift probably went
through an intermediate stage *r, and then assimilated to the medial *l. *d > *r > *l.
A related word (34a') may be inherited through another route. I posit that *d and *l
metathesized pre-PSek, and no assimilation took place, resulting in *kaladuŋ ‘deep’. Indeed, in
Seko Tengah, kaladuŋ and kadaluŋ are alternate forms with the same meaning. The variance
between *i in *i-laliŋ and *u in *kaladuŋ is unexplained in the final syllable.
(34)
a. *daləm
a'.
‘inside, deep’
*i-laliŋ
‘inside’
*ka-daluŋ, *ka-laduŋ ‘deep’
11
3.6
PMP *r
PMP *r became *∅ in PSek (35), so it seems to act like PMP *R (§3.10). Nevertheless, I am
treating it in its own section. Many PMP *r words have no traceable cognate in PSek.
(35)
3.7
PMP
PSek
*rəbuŋ
‘bamboo shoot’
*ohoŋ
*birit
‘rear end, buttocks’ *biiʔ
*baraŋ
‘perhaps, if’
*haŋ-kàliʔ
*garus, karus, gaut, ‘scratch, scrape’
*kauʔ SekP
kaRus, kaRud
PIn ?*ga(hø)u(k?) ‘work, do’ (Mills 1981)
‘shoot of palm or rattan’
‘perhaps, in case’
‘deed’
[semantic connection?]
PMP *Z
The pattern for PMP *Z is rather clear. It became *d in PSek word initially (36), and *r
word medially (37).
In (37e-g), *Z did not become *r medially because it was blocked by the nasal in the
consonant cluster. However, *nZ went to *nd before the nasal was lost, perhaps indicating that
*d was a transitional form in all *Z words that went to *r.
(36)
(37)
PMP
*Zalan
‘path, road’
*ZaRum
‘needle’
*Zaŋaw PWMP ‘paddy-bug’
PSek
*dalaŋ
*mu-dauŋ
*daŋo
‘sew’
‘stinkbug’
a.
b.
c.
d.
e.
f.
*uraŋ
*maN-turo:
*pa-taruŋ
*ka-rao:
*indaŋ
*tèdeʔ
[frozen prefix]
‘debt’
‘stand’
g.
3.8
*quZan
*tuZuq Zorc '71
*taZəm
*ma-Zauq
*hinZam PWMP
*tənZəg
*ti(n)Zəg
*inZak
‘rain’
‘point out’
‘sharp’
‘far’
‘borrow, lend’
‘upright, erect’
‘stand erect’
‘step on, tread on,
trample’
*idak SekP
PMP *g
Few Seko words reflect inherited PMP *g. Whatever there are in PMP become *k in PSek,
which has no *g.
(38)
PMP
*gəndaŋ Demp
*gənəp Demp
?*gunaq PIn Mills
?*gumaq < *Rumaq
‘drum’
‘complete, enough’
‘name’
‘sheath’
PSek
*kandaŋ
*kànaʔ
*kona:
*kuma:
Mmj guma
12
3.9
?*ga(m)bun Mills < *Rabun ‘cloud’
*kahuŋ
?*si(ŋ)gaq Mills
?*sugiR PIn
---
*mi-siŋka: SekP
*R > *ʔ irregular
*sukiʔ SekP
*aka
PSS *aga
‘stop by, visit’
‘rich’
‘what’
PSad *gawun
PMP *N
PMP *N is reflected as *n (39) in PSek in the very few correspondences that exist. Several
words with PAn *N already have doublets or another reconstructed form without *N in the lower
PMP node. Wherever PSek does not reflect PMP *N as *n, there is already another doublet that
it corresponds to. So in example (40), PSek *lite: corresponds to the doublet without the *N.
(39)
PMP
*ma-Najam PAn
PSek
‘accustomed to, tame’ *ma-naraŋ
(40)
*Nitəq, *litəq
(41)
*Nabuq Dahl, *nabuq ‘drop, fall’
3.10
PMP *R
‘sap of tree or plant’ *lite: SekP
*mi-naho:
‘smart, clever, skilled’
‘sticky sap’
‘slip and fall while
walking’
Most reflexes of PMP *R have become *∅ in PSek, whether initially (42), medially (43), or
finally (§3.14), such that exceptions present themselves as candidates for loans. For example
uraʔ at the end of (43) would be a loan.
(42)
(43)
PMP
*Rusuk Demp
*Ratus
‘rib’
‘hundred’
*maRi
*laRiw
*bəRas
*daRaq
*duRi
*ZaRum
*qasiRa
*suRuq Zorc '95
*bəRŋi
*bəRsay
*ma-Ruqanay
*uRat PWMP
‘come here’
‘flee’
‘rice’
‘blood’
‘thorn’
‘needle’
‘salt’
‘command, order’
‘night’
‘paddle’
‘male’
‘vein, root’
PSek
*osok
*atuʔ
*mai
*ma-lai
*heaʔ
*raa:
*dui
*mu-dauŋ
*sia
*maN-suo:
*heŋi
*(pi-)bose
*muane
*oaʔ
uraʔ SekP
‘rib, side’ [PSS *rusuk]
[PSS *ratus]
‘sew’
‘root, vein, tendon’
‘lines in palm of hand’
13
However, there are a few exceptions, and they merit mention. It’s hard to conceive of harani
being a loan in Seko Padang (44a), but the /r/ is unexpected. Perhaps *R > *r at some point
between PMP and PSek.6 The same goes for *kairi, *rea:, and *ma-ràba: (44b-d). Either they
were borrowed early on, or *R > *r in these words pre-PSek.
(44)
a.
b.
c.
d.
*baRanih PWMP
*ka-wiRi
*Riaq Zorc '95
*Rəbaq Zorc '71
3.11
PMP *w
‘brave, bold, dare’
‘left side’
‘sword grass’
‘destroy’
*harani SekP
‘bold, daring, savage’
*kairi
*rea:
*ma-ràba: SekP ‘broken’
PMP *w is quite straightforward, going to PSek *h (45). However, there are a few words
where the *w gets lost or is reflected as *u (46), of which I list all known reflexes. So, in (46a)
*w > *∅, in (46c) *w > *u, etc.
(45)
PMP
*sawa
*tawa
‘python’
‘laugh’
(46)
a. *wada
‘be, exist, have, wealthy’
b. *ka-wiRi
‘left side’
c. *wasay
‘axe’
d. *wahiR
‘water (fresh)’
e. ?*watɨR PIn
‘sago grub’
f. *quay
‘rattan’
g. *ka-wanan
‘right side’
h. *bawaŋ
‘open expanse of
land or water’
3.12
PMP *j
PSek
*saha
*ma-taha
*ara
*w > *∅
*kairi
*w > *∅
*uase
*w > *u
*uhai
*w > *uh
*uhati
*w > *uh
*uhe [*h suggests < ?*quway or ?*way]
*kuanaŋ
*a-w > *u
*hauŋ
‘river, stream’ *wa > *u
PMP *j went to PSek *r (47) via intermediate *d. *j > *d > *r. The shift *d > *r got blocked
in nasal clusters (48) just as it did for PMP *d and PMP *Z. And, *r assimilated to *l in the
environment of *l (49). Example (50) could be taken as a counterexample to the pattern, but
more likely it is indirectly inherited through a form without *j.
I’d like to mention two possible routes for the process of PMP *qaləjaw going to PSek *àlo
(49). Predictable consonant changes, if they all happened first, would have *qaləjaw > *aləlo.
Through predictable vowel changes, assuming regressive assimilation, we get *aləlo > *alàlo.
Dropping either *l plus a contiguous *a would leave *àlo. However, I don’t know of traces in
any language that support those intermediate steps. Alternatively, the process might have gone
through the following sequence: *qaləjaw > *aldo > *adlo > *àlo. That is more likely, based on
6
PKP *warani also exhibits an *R > *r irregularity in this word.
14
traces from existing languages, to wit, Ram ando, Kalao and Laiyolo ajo (< ?*alyo Mead
2003:134), Kapampangan aldo, Ceb adlaw. Mills (1975:696) reconstructed *ɨlzo for PSS.
PMP *j unexpectedly becomes *l in the cognate pair *ijuŋ > *ìloŋ (50a). But Seko is not
alone; several South Sulawesi languages and Bada' share this anomaly. The reflex begs for an
etymon like **idluŋ or **əluŋ.
(47)
PMP
*pija
*suja Zorc '95
*ma-dajəm
*pajay
*qapəju
*qujiŋ
‘how much?’
‘pitfall spike’
‘have chills’
‘rice plant’
‘gall (bladder), bile’
‘charcoal’
PSek
*pira
*sura
*mi-raruŋ SekP
*pare Bbd
*pùru
*oriŋ
‘how many?’
‘thorn’
‘feel cold’
‘year’ Psn
[*a > *∅]
(48)
*li(n)juq Mills ‘front, forehead’
*lindo:
*huaji, *ha(n)ji Mills ‘younger sibling’ *adi
(49)
*qaləjaw
*laja Demp
‘day’
‘burn (a wound)
be hot (spices)’
*àlo
*mi-lalaŋ SekP [*∅ > ŋ, perhaps a loan]
‘ginger’ Psn, Bdb
(50)
*ijuŋ
*iloŋ PSS
‘nose’
*ìloŋ
[expected < ?*əluŋ or ?*idluŋ]
illoŋ in Dur, Klp, Rkg
3.13
‘forehead’
Nasal Clusters
Sequences of nasal plus stop in PMP are retained in PSek, with the nasal assimilated in point
of articulation to the following stop. All nasal clusters in PSek are word-medial. No words begin
with a pre-nasalized consonant. For an illustration of what happens to PMP non-nasal clusters,
see §4.2.1.3.
(51)
3.14
PMP
*ampu
‘grandparent, grandchild’
*kambaŋ
‘bloom, well up, swelling’
*punti
‘banana’
*diŋdiŋ
‘wall’
*andap PWMP ‘luminous millipede’
*laŋkaq
‘step, stride’
PSek
*ampe
*kambaŋ
*punti
*rindiŋ
*andaʔ
*laŋka:-laŋka:
[unexplained *u > *e]
‘swollen’
‘spider’
Final Consonants
All the word-final consonants in PMP are reduced to three in PSek, plus one prosodic
feature of length on final vowels, as illustrated in Table 2. The rest have been dropped. PMP *-l
has reflexes of *ŋ, *l, and *∅, but without a clear pattern. Final *-w and *-y are treated under
diphthongs (§4.1).
15
Table 2. Proto Seko word-final consonants
PMP
PSek
(52)
*p, t, b, d, g, s, j, r?
*ʔ
PMP
*qatəp
*laŋit
*-b
*sabuD
*ma-panas
*palaj
*ənəm
*quZan
*katəl
*tasak
*puluq
*buaq
*m, n, ŋ, l
*ŋ
‘roof’
‘sky’
*tuqah
*pul Zorc
*bəlbəl
*qatəluR
*naquR PPh
‘sow, scatter’
‘warm, hot’
‘palm, sole’
‘six’
‘rain’
‘itch’
‘ripe’
‘ten’
‘fruit, areca palm
and nut’
‘old’
‘blunt, dull’
‘plugged, stupid’
‘egg’
‘descend’
*-r
*ilah
‘wild’
*k
*k
*q, l?
*:
*l, R, r?, h
*∅
PSek
*atuʔ
*laŋiʔ
no known correspondence
*maN-sahuʔ
*ma-panaʔ
‘sick, in pain’
*palaʔ
*ùnuŋ
*uraŋ
*mi-katiŋ
*tasak
‘done cooking’
*pulo:
*hoa:
‘fruit’
*bua:
‘areca palm and nut’
*ma-tua:
[metathesis]
*pompo:
*bebe SekP
‘mute, stupid’
*tùlu
*nau:
‘go downhill/
downstream’
no known correspondence
*ma-ila
The only known sources of -r > -ʔ are loans, like pasa' ‘market’ < pasar (Mal) and kolla
pasi' ‘granulated sugar’ < gula pasir (Mal). If the latter were inherited from PMP *pasiR, we
should get pasi, but we do not. It is my guess that PMP *-r became *-∅, like *-R did, and not
*-ʔ, like in loans, but so far I have not found any clear cognate pairs.
There are a couple of words where *-R unexpectedly became -ʔ in SekP (53). damaʔ ‘resin’
is probably a loan from Malay damar, from when its economic importance was introduced.
Because of the glottal, ma-tahaʔ could be classified a loan, but it can also be explained by
differentiation to avoid a homonym, because ma-taha already means ‘laugh’.
(53)
*damaR Zorc '95 ‘resin’
*tabaR
‘flat, tasteless’
damaʔ
ma-tahaʔ
[expected dama]
[expected ma-taha]
16
4.
Vowels, from PMP to Proto Seko
4.1
Regular changes
The regular changes of vowels and diphthongs are presented in Table 3, followed by
examples (54). PMP *ə will be treated separately, beginning in §4.2.2.
Table 3. Unconditioned sources for Proto Seko vowels and diphthongs
PMP
PSek
(54)
PMP
*taliŋa
*buni
*pitu
*kahiw
*baliw
*pisaw
*matay
*babuy
*i
*i
*a
*a
*u
*u
*-iw
*-i
‘ear’
‘hide’
‘seven’
‘wood’
‘dual division, moiety’
‘knife’
‘die’
‘pig’
4.2
Conditioned vowel changes
4.2.1
Vowel lowering
*-aw
*-o
*-ay *-uy
*-e
*-i
PSek
*taliŋa
*miN-ka-buni (vi), maN-huni (vt)
*pitu
*kayu
[rule ordering applies]
*hali
‘in balance’
*piso
‘machete’
*mate
*bahi
When I refer to vowel lowering, I mean to account for PSek reflexes *e and *o, inherited
from PMP *i and *u respectively, beyond what was already described in §4.1. There are many
/o/s and /e/s in Seko languages, and it’s difficult to account for them all, but I will describe what
patterns I can. We will start with one of the clearer conditioning environments for vowel
lowering, PMP *-q.
4.2.1.1 Lowering before *-q
We saw in §3.14 how PMP *-q is reflected as length on final vowels in Seko languages. In
addition to length, high vowels get lowered, with the result that PMP *-iq > *-e: and *-uq > *-o:
(55).
(55)
*uliq ‘return home, return something; restore,
repair; repeat; motion to and fro’
*pusuq
‘heart’
*tu(m)buq
‘sprout, grow’
*bunuq
‘kill’
*qipuq
‘tree with poison sap’
*ule:
*puso:
*tuho:
*buno:
*ipo:
‘live, grow’
‘take heads’
17
4.2.1.2 Lowering of *u before *-ŋ
PMP *u became *o in PSek before PMP *-ŋ, but not before PMP *-m or *-n.7
(56)
*ijuŋ
*əsuŋ PWMP
*rəbuŋ
*lutuŋ
*ZaRum
*dahun
‘nose’
‘rice mortar’
‘bamboo shoot’
‘monkey’
‘needle’
‘leaf’
*ìloŋ
*ìsoŋ
*ohoŋ
*ma-lotoŋ
*mu-dauŋ
*dauŋ
‘shoot of palm or rattan’
‘black’
‘sew’
4.2.1.3 Lowering in reduplicated monosyllabic roots
A PMP C1C2 consonant cluster in reduplicated monosyllabic roots becomes C2 in PSek,
provided C1 is not a nasal. Nasal consonants are retained. Often the preceding vowel becomes
lowered and/or sometimes takes on the quickness feature.
(57)
*kitkit
?*kuhkuh
*tuktuk
*bukbuk
?*maN-bukbuk
*suksuk
*pulpul Zorc
*maN-kekiʔ
*maN-koko SekP ‘reach inside
and take, or fondle’
‘knock, pound, beat’ *maN-tùtuk SekP ‘pound into pieces’
‘weevil, dust
*bubuk
SekP ‘frass, leavings’
produced by weevil’
?‘bore’
*maN-bobok SekP ‘dig horizontally’
‘stab’
*maN-sosok SekP ‘insert, circulate’
[cognate?]
‘blunt, dull’
*pompo:
‘bite’
‘scrape’
4.2.1.4 Lowering word initially in the environment of *R
Vowels in PSek have been lowered in word initial position where contiguous to a PMP *R.
This happens in only two or three words, but there appear to be no counterexamples. *R does not
lower vowels in other word environments. While we cannot really say that *ə is ‘lowered’ to *o
in (58c), neither does it completely harmonize with *u (see §4.2.2). It is reasonable to ask
whether *-k might be the determining factor, but there are counterexamples where *-uk is not
lowered. Since the evidence is slight, it might even be better to conclude that these examples
only represent spontaneous innovation.
(58)
a.
b.
c.
*Rusuk
*uRat
*rəbuŋ
‘rib’
‘vein, tendon’
‘bamboo shoot’
*osok
*oaʔ
*ohoŋ
‘rib, side’
also, ‘root, vine’
‘shoot of palm or rattan’
4.2.1.5 Lowering in the environment of PMP *ə
Independent of how *ə gets reflected segmentally in PSek, PMP *i and *u become lowered
when followed by *ə. This is height assimilation.
7
This analysis is tentative, however, and not without counter-examples, like PMP *patuŋ ‘large, thick, bamboo sp’
> PSek *pantuŋ. I haven’t had opportunity to investigate whether this might be a loan.
18
In (59e), *toaʔ could reflect either of the two etyma, but *tuqəD is preferable because the
*-əR- combination in other words is usually reflected as *-e- (§4.3.3.4). If that applied here,
*təRas would be reflected **teaʔ.
(59)
4.2.2
a.
b.
*inəm
*ipəs
‘drink’
‘cockroach’
c.
d.
e.
*kudən
*bituqən
*tuqəD
*təRas
‘cooking pot’
‘star’
‘stump’
‘heartwood’
*m-enuŋ
*epiʔ SekP
*i > *e
*ə > *u
*i > *e
*ə > *i
‘small cockroach sp’
*koriŋ
*u > *o
*ə > *i
*bintoeŋ
*u > *o
*ə > *e
*u > *o
*ə > *a
*toaʔ
‘trunk, base (of tree)’, Mal pangkal
Vowel harmony in reflexes of PMP *ə
Besides vowel lowering, the only conditioned vowel change I am aware of for PSek is
vowel harmony. In this set of reflexes, which has well-attested correspondences, the PSek reflex
of *ə harmonizes with the other vowel in the word (60). It appears that the neighboring vowel is
a strong conditioning factor. Notice that the reflex in a penultimate (stressed) syllable has
quickness, marked by a grave accent.
(60)
4.3
PMP
*təbuh
*qatəluR
*bənaŋ
*təlaŋ PWMP
*baləs
‘sugar cane’
‘egg’
‘thread’
‘bamboo sp’
‘reciprocate’
PSek
*tùhu
*tùlu
*bànaŋ
*tàlaŋ
*maN-balaʔ-i
maN-balaʔ
*bìtiʔ
*pùno:
*teme:
*bətis
*pənuq
*kəmiq
‘calf of leg’
‘full’
‘urinate’
*sədut PWMP
*qapəju
*si(n)jəm
‘sip, suck’
*sùruʔ
‘gall (bladder), bile’ *pùru
‘ant sp’
*siriŋ SekP
‘buy’
‘urine’
[no quickness]
[*k > *t unexplained]
‘sip, suck, slurp, inhale’
Unconditioned vowel changes
The rest of the vowel changes from PMP to PSek are unconditioned, as far as I can tell,
although I suggest some possibilities related to conditioning that could be investigated further.
Perhaps the reader can pick out a pattern that has escaped the author’s notice. Sporadic vowel
lowering (§4.3.1) is a tendency, and most of the other changes are the various reflexes of *ə not
following the pattern of vowel harmony presented in §4.2.2. Each reflex of *ə will be presented
in its own section. I try to limit my discussion to patterns, and place comments relating to
specific etyma under their reconstructed forms in §7.
19
4.3.1
Unconditioned vowel lowering
4.3.1.1 PMP *u > PSek *o
Vowel lowering in (61a) and (61b) should not be explained by *q- and *-q-, for they are not
lowering environments for PSek, and there are several counterexamples. I am inclined to suggest
reconstructing the second vowel in those etyma to *ə for a smoother reflection of the PSek, i.e.,
?*qujəŋ (dbl *qajəŋ already reconstructed, Blust ACD) and ?*quqəs. Such reconstructions would
allow those correspondences to be characterized as having ‘conditioned vowel changes’
(§4.2.1.5)
Reconstructing PMP *kukud as ?*kudkud (see §4.2.1.3) or ?*kukəd would also allow this
cognate pair (61c) to be characterized as having a ‘conditioned vowel change’.
In (61d), *u could be lowered due to the presence of *-q (see §4.2.1.1), even through they
are not contiguous, and the same for *i in (62a).
PMP
a. *qujiŋ
‘charcoal’
b. *ququs
‘chew sugarcane’
c. *kukud
‘shank or hoof of animal’
d. *luaq Dyen
‘expel from mouth’
(61)
PSek
*oriŋ
*mu-oiʔ
*kokoʔ
*loa:
‘leg and foot’
‘spit out s.t.’
4.3.1.2 PMP *i > PSek *e
Etyma in (62) have *i preceding *u. This could be a pattern of dissimilation of high vowels.
(62)
a.
b.
c.
*hinuq
*iluR
*inum
‘beads’
‘saliva’
‘drink’
*eno: SekP
*elu SekP
*enuŋ
‘necklace’
And the lowering continues. Among Seko languages, PSek *u in *manuk ‘chicken’ is
reflected as /u/ in Panasuan, but as /o/ in the other three. *tau ‘person’ is contracted to an
unstressed, prefixal form to in Seko Tengah, but tu in the other three. Intransitive verbal prefixes
mu-, mi-, miN-, and ti- have become mo-, me-, meN-, and te- in Seko Tengah, but then the pattern
of vowel reduction is only natural in pre-penultimate position (Sirk, p 60).
4.3.2
Sporadic changes from PMP *i, *a, and *u
PSek *tìtiʔ would fit better if PMP *titis (63) were reconstructed ?*tətis, ?*təstəs, or ?*tistis.
(63)
*titis
‘drip’
*tìtiʔ
‘drop’
[*i > *ì ]
Vowels in the antepenultimate position can be ‘weak’ and easily influenced by their
environment. That might be why *a > *u in (64), due to the influence of *b.
(64)
*balaqih
‘co-parents-in-law’
*bulai
In Seko, *ampe is distinct from *ampu (65a). *u > *e is unexplained.
[*a > *u]
20
(65)
a. dbl *əmpu
a'. *əmpu
b. *hutək
‘grandparent, grandchild’ *ampe
‘owner’
*ampu
‘marrow, brain’
*etik
[*u > *e]
[*u > *e]
4.3.3 Unconditioned reflexes of PMP *ə
In this section, unconditioned reflexes of etyma having PMP *ə are presented, a separate
subsection for each PSek reflex *a, *i, *u, *e, and *o. Conditioned reflexes which follow the
pattern of vowel harmony were already presented in §4.2.2.
*ə is the mother chameleon of Proto Austronesian. Its reflexes blend into the phonological
environment to the extent that it is difficult to even say when it is involved in an unconditioned
sound change. Patterns do emerge, but are not as consistent as one might hope, and I prefer to
call them tendencies. For example, *ə tends toward being reflected as *a in PSek, but the other
vowels are also represented. Perhaps some tendencies will reveal secondary conditioning factors
to the observant.
Keep in mind that a PSek reflex of *ə in a penultimate (stressed) syllable usually has
quickness, marked with a grave accent.
4.3.3.1 PMP *ə > *a
(66)
*gənəp Demp
‘complete, enough’
*qənəp CEMP ‘lie down to sleep’
*qənəŋ PMP ‘quiet, still, at rest’
*təktək
‘cut (wood)’
*bətik
‘tattoo’
*ma-əmis
‘sweet (taste)’
*təlu
‘three’
*əmpu
‘grandparent, grandchild’
*əmpu
‘owner’
*tuqəD
‘tree stump, stubble’
*kànaʔ
*m-ànaʔ
*tatak SekP
*bàtik SekP
*màmiʔ
*i-tàlu
*ampe
*ampu
*toaʔ
‘quiet, still, at rest’
[ŋ > ʔ irregular]
‘chop up for cooking’
‘speckled, blotched’
‘delicious’
‘base of tree’
4.3.3.2 PMP *ə > *i
*s could be a conditioning environment for PMP *ə becoming *i in PSek. Other possible
consonants to consider as conditioning environments are *l, *t, and *j.
(67)
*sisiʔ
*pìse:
*sali
*mi-katiŋ
*i-laliŋ
*daliʔ
*m-eiʔ
*epiʔ SekP
*səpsəp
*pəsəq Zorc
*saləR
*katəl
*daləm
*laləj
*ikəj PWMP
*ipəs
‘suck’
‘break, broken’
‘floor’
‘itch’
‘inside, deep’
n ‘fly’
‘cough’
‘cockroach’
*litəq
*lisəhaq
*kudən
‘sap of tree or plant’ *lite: SekP
‘nit, louse egg’
*lise:
‘cooking pot’
*koriŋ
‘suck on’
‘inside’
[expected *laliʔ]
*i > *e
*ə > *i
‘k.o. small cockroach’
‘sticky sap’
21
*hutək
*əsuŋ PWMP
*əjuŋ
‘marrow, brain’
‘rice mortar’
‘nose’
*etik
*ìsoŋ
*ìloŋ
4.3.3.3 PMP *ə > *u
The reflexes in (68) indicate that PMP *e in the environment of *m or *p > PSek *u.
(68)
*ənəm
*əpat
*daləm
*ma-dajəm
*taZəm
*adəp PWMP
*qatəp
*inəm
‘six’
‘four’
‘inside, deep’
‘have chills’
‘sharp’
‘chest’
‘roof’
‘drink’
*ùnuŋ
*ùpaʔ
*ka-laduŋ
*mi-raruŋ
*pa-taruŋ
*aruʔ
*atuʔ
*maN-enuŋ
‘deep’
‘feel cold’
4.3.3.4 PMP *ə > *e
(69)
*bəlbəl
*tənZəg
*bəRas
*əZan
*bəRŋi
*bənut
*bituqən
‘plugged, stupid’
‘upright, erect’
‘rice between harvesting
and cooking’
‘notched log ladder’
‘night’
‘husk’
‘star’
*bebe SekP
*tèdeʔ
*heaʔ
‘mute, stupid’
‘stand’
*ereŋ SekP
*heŋi
*benuʔ
*bintoeŋ
‘ladder, stairs’
‘pierce, stab’
‘paddle’
‘bamboo shoot’
*maN-tòbok
*(pi-)bose
*ohoŋ
‘poke’
4.3.3.5 PMP *ə > *o
(70)
4.4
*təbək
*bəRsay
*rəbuŋ
‘shoot of palm or rattan’
Evaluation of PMP *ə
The complexity in the historical analysis required to account for correspondences of *ə are
disproportionally high in Seko compared to other issues. This seems unwarranted, and causes me
to seek an explanation. Two possibilities for the complexity come to mind. The most likely is
that, unlike Seko and South Sulawesi languages, many other languages have lost the contrasts
necessary to directly reflect PMP *ə, such that etyma with *ə are more in need of revision than
are etyma without *ə. The seems to be the source of most of Proto Seko’s irregularities.
The second possibility lies in the nature of PAn *ə itself. In order to explore that issue, let’s
address the question, what is it that PAn *ə contributes to the phonology that brings about such
complexity to the reflexes in PSek, and other languages as well? And what does PAn *ə have to
do with vowel quickness in PSek? The answer that I think the Seko data shows us is that PAn *ə
carried a prosodic feature that is normally associated with closed syllables.
22
In fact, a possibility that toys at my mind is whether *ə should be analyzed as a segmental
phoneme at all. Rather, it could be viewed as a prosodic feature superimposed on the other three
vowels *i, *a, and *u. I’m not going to go there, however. I can’t see the way out, and it is likely
to be a fruitless pursuit. Let’s just assume that PAn *ə was a segmental phoneme, but I would
add that it differed from the other vowels in that it had an associated prosodic feature,
characterized by a truncated timing mora or absence of a coda blocker. What besides prosody
could have possibly motivated formation of quick vowels (ambisyllabic consonants) in PSek,
geminate consonants in South Sulawesi languages, schwa plus stress shift in Malay, and at least
a portion of contrastive stress in Philippine languages. I cannot imagine that a simple segmental
phoneme is capable of doing all that.
4.5
Conclusions about Rule Ordering from PMP to Proto Seko
Our story about the sound changes from PMP to PSek would not be complete if we did not
arrange them in a proposed chronological order. In Figure 2, tie bars connect ordered sound
changes. There are a few minor internal conflicts between the data for supporting one order over
against another, which cannot be resolved without further research. I’m sure the sequence will
need to be reordered somewhat as genetic relationships are clarified for Sulawesi languages in
time to come. Example (71) illustrates how rule ordering applies for PSek.
Consider the following subset of ordered rules, taken from Figure 2:
*R
*-iq, uq
*-qV
*-q*-q
*-C
(71)
*puluq ‘ten’
>
*t-in-aqi ‘guts’
>
*naquR ‘descend’ >
*laŋit
‘sky’
>
> *∅
> *eq, oq
> *-Vq
> *∅
> *-:
> *-q8
*pulo:
*tai:
*nau:
*laŋiq
Loss
Lowering
Metathesis
Loss
Vowel length
Merger
not *pulu:
not *tai or *tae:
not *nau or *nao:
not *laŋi:
In another sequence of sound changes, it appears that a proto-form *kanahaw ‘sugar palm’
(PWMP *qanahaw) > *kanaaw > *kanaau > *kanaqu > *kanauq > *kanau: (72). The crucial
timing in this sequence is that the shift from -aaw to -aqu must have happened before the *-aw >
*-o rule could apply (see Figure 2).
(72)
dbl ?*kanahaw ‘sugar palm’ > *kanau:
not *kanao
Examples in (73) were very possibly inherited indirectly through neighboring Central
Sulawesi languages. They would have been borrowed after *-uq > *-oq, but before *-q > *-:.
q = glottal, here and in *laŋiq (71) and *walesuq (73). It is written *ʔ elsewhere in this paper for PSek, and in PKP
it is written *'.
8
23
(73)
no etymon found ‘rat, mouse’
*suluq
‘torch’
X
balesu:
sulu: SekP
[PKP *walesuq]
[expected sulo:]
X
Figure 2. Ordered sound changes, PMP to Proto Seko
*V+high
*-C1C2*h
*-aaw
*-aw, -ay
*-iw, -uy
*R
*V+high
*R, r
*aə
*ə
*-iq, uq
*k-, -k*-qV
*-V
*d*j
*q-, -q*ñ
*N
*g
*b
*w
*Z
*-d*d, r
*NC
*u
*-m, -n
*-q
*-C
> *V-high / C1C2 (C1 ≠ nasal)
> *C2 (C1 ≠ nasal)
> *∅
> *-aqu
> *o, e
> *i
> *g few
> *V-high / #R_, #_R
> *∅
> *a
> *a, i, u, e, o
> *eq, oq
> *∅ partial
> *-Vq
> *-VN / _-POSS
> *r- partial
> *d
> *∅
> *n
> *n
> *k
> *w partial
> *h, u
> *d
> *-r- / V_
> *l
/ l partial
> *C partial
> *o
/ _ -ŋ
> *-ŋ
> *-:
> *-ʔ
Lowering / Redup
Deletion
Reanalysis
Lowering / R
Harmonization, etc
Lowering / -q
Metathesis
Possessive form set
Lowering / -ŋ
Merger
Vowel length
Merger
PSek *sumpi: (74) was probably also borrowed, but from an unknown source. *sumpit >
*sumpiq > *sumpi:. Cf PKP *supi'.
(74)
*su(m)pit
‘tongs’
X
sumpi:
[expected > **sumpiʔ]
24
5.
From Proto Seko to the Present
How do the Seko languages today reflect Proto Seko? How are they related to one another?
Table 4 presents the innovations.
Table 4. Phonological innovations within the Seko group
Language - Dialect
Seko Padang
Seko Padang
Seko Padang - Hono'
Proto Seko
*-Vαk
*nd, mb
*l
Seko Padang, Seko
Tengah - Ambàlong
Seko Tengah
Seko Tengah
Panasuan,
Budong-budong
Panasuan,
Budong-budong
Panasuan
Panasuan
*-kiŋ suffix
Innovation
-VαkVα
nn, mm
ɭ / V-front_V+front
-iŋ
*mp, nt, ŋk
*i, u in prefixes
*-k
pp, tt, kk
e, o
-ʔ
Example
*anak
> anaka ‘child’
*rindiŋ > rinniŋ ‘wall’
*ale:
> [ aɭe: ] ‘mat’
*maN-rène'-kiŋ > maN-rène'-iŋ
‘drop’
*kuntuʔ > kuttuʔ
‘knee’
*mi-kehiʔ > me-kehiʔ ‘hot’
*anak
> anaʔ
‘child’
*r
h
*raa:
*d
*t
Budong-budong
Budong-budong
*V:
*y
r
tʃ / _V+high
V often no length
dʒ
*adi
> ari ‘younger sibling’
*pitu
> pitʃu
‘seven’
*(ma-)sàle: > ma-sàle ‘cold’
*kayu
> kadʒu
‘wood’
Budong-budong
*-a
-ɔ sometimes
*mi-taha
> haa:
‘blood’
> mi-tahɔ ‘laugh’
Several of the innovations in Seko languages reflect areal influences. Those are:
1. Retroflection of *l as an allophone in the Hono' dialect of Seko Padang. Rampi, Bada',
Behoa, and Uma have retroflexed l allophones, too (Martens 1990:42).
2. The Seko Tengah practice of taking nasal - voiceless stop clusters and turning them into
geminate voiceless stops follows the practice of Kalumpang.
3. Most surrounding languages have lowered vowels in pre-penultimate prefixes: Uma,
Rampi, Bada', Rongkong, Kalumpang, Mamuju, Tabulahan, Bambam.
4. Final k has become glottal stop or been lost in Karama dialects of Kalumpang, but
retained in the Bone Hau dialect.
5. *r > h in Bambam, Tabulahan, and Talondo', but not in another Bambam dialect and not
in Kalumpang.
6. In recent generations, Budong-budong has been isolated from other Seko languages.
Their current location has brought them into close contact with Topoiyo and Mamuju.
25
Their numbers are small and the influence of those languages on Budong-budong is high.
Besides many lexical replacements, the following areal influences can be observed: loss
of final vowel length, fortition of *y > dʒ (also > nʲ), and backing of *-a > -ɔ (probably
from Topoiyo).
Figure 2. The Seko group relationships based on sound change
Proto Seko
*r > h
*-k > ʔ
Proto Panasuan
*: > ∅ (partial loss)
*-a > -ɔ (partial)
*y > dʒ
*mp > pp
*nt > tt
*-king > -ing (suffix)
*mb > mm
*nd > nn
*-Vαk > -VαkVα (echo
vowel after -k)
*ŋk > kk
Seko Padang
*d > r
*t > tʃ
(before V+high)
Panasuan
Budong-budong
*i > e, *u > o
(in prefixes)
*l > l, ɭ (retroflexed
allophone)
Seko Tengah
Lodang
Hono'
Notice that one ordering rule is important for Panasuan. That is, *r > h before *d > r.
Otherwise we would find words where *d > h, but we do not.
6.
Toward classification: comparison to Sirk’s criteria for a South Sulawesi group
Now we may ask, where does Seko fit in the grand scheme of things? I am not prepared to
go into that deeply, but let me offer a first step.
In a 1989 article, Sirk set forth eight criteria that he said characterized the South Sulawesi
group (SSul). Against the framework of those criteria, we compare four groups in Table 5,
namely, Tamanic (Adelaar 1992), PSS (Mills 1975), PSek, and Badaic9 (Martens 1990).
9
Martens reconstructed Badaic and Rampi as a subgroup of Proto Kaili-Pamona, but Mead (2003) stated they did
not belong there based on reflexes of PMP *j (pp 131-133), and the lowering of *-uq > -oq (p 136).
26
Table 5. Comparison to Sirk’s SSul group criteria
Characteristic of SSul
Tamanic
1. PMP *-q > *-q
2. nasal ligature
before -POSS
3. *R > *r
4. *-C2C1-10 > *-ʔ C1(C2 ≠ N)
5. Transitive
prefix *uN6. Numerals ‘eight’
and ‘nine’
yes -ʔ
no
r
--?
borrowed
borrowed
sera ‘one’
7. Lexical items group I
--8. Lexical items group II
--********
*****
9. PMP *-uq > *-oq
yes
(vowel lowering)
10. PMP *əCV >
aCV
*təbuh ‘sugarcane’ >
---
tabu
sadiŋ
‘wake up’
PSS
no *-∅
yes
PSek
yes *-:
yes
*r
*-ʔ C1yes
yes *-ʔ
no
∅
∅
*-C1-
*-C1-
no
*karua(a)
*karoaa:
*kamesa(a) / *kamesaa:
*kasera(a)
16 items
5/16 cognate
12 items
4/12 cognate
*****
*****
yes
yes
*əCCV
*àCV, etc
(gemination) (quick vowel)
*təbbu
*səddiŋ
‘feel, hear’
Badaic
*tùhu
*sàdiŋ
‘hear, feel’
no
*u-walu
*sa-sio
----*****
yes
VCV
tuwu
hádi
‘hear, feel’
We can see in Table 5 that Sirk took issue with Mills’ conclusion that PSS did not reflect
PMP *-q. Mills was correct, in that none of the other South Sulawesi languages directly reflect
*-q, but Seko languages, Badaic, and Tamanic do.
Regarding items 2-8, Sirk is in harmony with Mills for PSS, but the other three language
groups do not line up as well. In comparison with PSS on sound changes, only PSek has nasal
ligature and only Tamanic reflects *R as r. Neither PSek nor Badaic reflects the first consonant
(C2) in earlier reduplicated monosyllabic roots.
Items 5-8 are lexical. PSek agrees nicely with PSS on the numerals, but the tally for other
lexical correspondences is quite low.
I have added two more points of comparison to those of Sirk. All four groups of languages
lower PMP *-uq to -oq, something Mead (2003:136) considers diagnostic. Item 10 shows the
various ways these language groups handle the PMP sequence *əCV. In PSS, geminate
consonants are the outcome, and in PSek quick vowels. Tamanic seems to consistently reflect *ə
as a with no other segmental change. Badaic maintains a single intervocalic consonant, but the
reflex of *ə may harmonize with the other vowel, as in tuwu. It appears to have lost the closed
10
The full symbolization for proto reduplicated monosyllabic roots is *C1VC2C1VC2.
27
syllable prosody associated with *ə, but on the other hand, a trace may remain. For, Bada' and
Behoa have an unrounded, low, back vowel [ ɑ ], spelled á, which corresponds in this word.
My conclusion from Table 5 is that the four language groups are definitely related, but that it
is unlikely the other three are inherited from PSS, even if it were revised to reflect PMP *-q.
How the nodes are related will take further investigation. The key, I believe, to further
understanding the relationship of Proto Seko to these other language groups will be in accurately
applying ordered sound change rules, such as those shown in Figure 2. For sure, the definition of
PSS will have to be changed. It should either be expanded and reworked to include more
languages, or reduced and tightened up after removing Seko languages at least. In the latter case,
a new term will be needed to cover the broader range.
7.
Lexical Reconstructions
7.1
Introduction
The reconstructions presented in §7.2 are drawn from 488-word lists I collected from all
dialects of the Seko group and surrounding languages. Some of the lists for neighboring
languages, taken for comparison, were collected by SIL International colleagues. Section 7.3 is
an additional wordlist based solely on Seko Padang data from my own fieldwork, for which
reconstruction is rather straightforward or certain.
Reconstructed entries are structured as follows. On the first line is the reconstructed Proto
Seko etymon. This is followed by an English gloss, and sometimes, for clarification, an
Indonesian gloss. The gloss is sometimes followed by a question mark and a code letter, meaning
that the link between the PSek reconstruction and the PMP etymon is in doubt. ?d means the
distribution is limited. ?p means a heavy phonological irregularity. ?s means a tenuous semantic
correspondence.
The second line begins with identification of which languages reflect the reconstruction
regularly (blue font). Following a semicolon, reflexes are cited for languages where they deviate
unpredictably from the PSek form. Next, in brackets, the level and etymon are cited from which
the PSek reconstruction is inherited, followed by the source in parentheses. Following the closed
bracket are comments about irregularities, expected reflexes or etyma, possible pathways of
inheritance, and related forms in other languages.
I have used the following abbreviations for language names in the Seko group.
Abbreviations for other languages referred to are footnoted.11
Seko Padang
Seko Tengah
11
SekP
Hono'
Lodang
SekPH
SekPL
Ambàlong
Hoyane
Pokappaang
SekTA
SekTH
SekTP
SekT
Bad Bada', Bam Bambam, Beh Behoa (Besoa), Blt Balantak, Ceb Cebuano, Dur Duri, Enr Enrekang, Gor
Gorontalo, Klp Kalumpang, KlpBH Kalumpang - Bone Hau, KlpKr Kalumpang - Karama, Klw Kulawi, Kul
Kulisusu, Mak Makas(s)ar, Mdr Mandar, Mmj Mamuju, Mms Mamasa, MorB Mori Bawah, Nap Napu, Pen Pendau,
Ram Rampi, Rkg Rongkong, Sad Toraja Sa'dan, Skt Sanskrit, Tab Tabulahan, Tae Tae', Tam Tamanic dialects, Tld
Talondo', Top Topoiyo
28
Panasuan
Budong-budong
7.2
Psn
Bdb
Proto Seko lexical reconstructions
Entries preceded by superscripted X reference loans.
In expected etyma, -C represents an unknown proto consonant from the set (*p, t, b, d, g, s,
j, r) from which PSek glottal (ʔ ) is regularly reflected.
*adi ‘younger sibling’ (possessive suffix required), *adi: (vocative)
All [PMP *ha(n)ji Mills 616)]12 regular if *nj > *nd > *d
*i-aho ‘above, on top of’
All [PMP *i-babaw (Blust 1999)] presumed *b- > *h- > *∅*aka ‘what?’
All [possibly a mutation of PMP *apa (Blust 1999)] cf PSS *aga; only Bugis reflected as
g, PUS dialects and KlpBH as k
*ale: ‘mat’
SekP, SekT, Psn. Suggests < ?*aliq. Cf PKP *ali', PPUS *ale, PSS ?*jali(C?). PMP
*aliq ‘move, change place’ (Dempwolff, in Zorc 1971 item 0109) has a semantic
connection (mats are portable), but it seems too distant
*àlo ‘day’
All [PMP *qaləjaw (Blust 1999)] *j > *r > *l, then *qal- > *∅-. Or, *ə > *∅, then *lj >
*ld > *dl > *l. Cf Ram ando, (aldo Martens pers. com.)
*alu ‘rice pestle’
All [PMP *qahəlu (Blust ACD)]
*ambuŋ ‘fog, mist’
SekT, Bdb; SekT also, ‘cloud’ [PWMP *ambun ‘dew, mist, fog’ (Blust ACD)]
*ampe ‘grandparent, grandchild’
SekP, SekT [PMP dbl *əmpu (Blust ACD)] *u > *e unexplained, appears to be an
innovation. Cf PPUS *ampo ‘grandchild’, PSS *ɨmpu ‘grandparent, grandchild’
*maN-ampe-i ‘wait for’
SekP, Psn; Bdb maN-empe-i, SekTH maN-appo-i (some vowel irregularities) [PMP
*qantay (Dempwolff 1938)] *nt > *mp unexplained. Cf Tab maN:-empe, Mmj maNeppe, Bam maN:-ampa, KlpKr ma'-kappa, Rkg makampa. Together these suggest <
?*kəmpay
*ampu ‘owner, citizen, resident, host, property, head (as of a clan)’
SekP; Bdb ampu hai ‘crocodile’ [PMP *əmpu ‘lord, master, owner’ (Blust ACD)]
12
Blust (ACD) reconstructs PMP *huaji, with no comment on the uncertain (ń) in Dempwolff’s *a(ń)ji. Since the
presence of *n in the etymon fits best with PSek, I’m not sure how to reconcile with today’s upper level
reconstructions.
29
*anak ‘child’
All [PMP *anak (Blust 1999)]
*andaʔ ‘luminous millipede’
All [PWMP *andap (Blust 2002)]
*ande: ‘eat’ (see *kande:)
*anisi ‘tooth’
SekT, Psn; Bdb nisi [PMP dbl *ŋisi ‘grin, tooth, gum’ (Blust ACD)] uncertain whether
PSek *a- is a regular development preceding PMP *ŋ*api ‘fire’
All [PMP *hapuy (Blust 1999)]
*aruʔ ‘chest’
SekP [PWMP *adəp (Blust ACD)] appears related to SekT mabaru', Psn, Bdb babahu'.
*asu ‘dog’
All [PMP *asu (Blust 1999)]
*atuʔ 1 ‘roof’
All [PMP *qatəp (Blust 1999)]
*atuʔ 2 ‘hundred’
All [PMP *Ratus (Dempwolff 1938, Dyen 1953, Zorc 1995)] cf PSS *ratus (Mills
1975:299)
*baak ‘head’
SekT, Psn, Bdb; SekP baka (expected baaka). Cf PSad ?*baka13, PMAT *baak, Bad
woaq, Uma wooq, Bar woqo (Mills 627)
*baba: ‘doorway’
SekP, SekT; Bdb baba-ang [PMP *baqbaq ‘mouth’ (Blust 1999)]
*maN-babe ‘make (a mat, e.g.)’
All. Cf Uma, Nap babehi, Behoa bábehi, possible frozen suffix -hi
*bahi ‘pig’
All [PMP *babuy (Blust ACD)]
*maN-balaʔ ‘buy’
SekP; SekP maN-balaʔ-i ‘repay, reciprocate’, miN-balaʔ ‘retaliate’ [PMP *baləs ‘answer,
retaliate, reciprocate good or evil’ (Blust ACD)]
*balambaŋ ‘belly’
SekP, SekT. Frozen prefix? ba-, although frozen infix -al- might also be possible. Cf PSS
*balɨmba(N)/barɨmba(N) ‘chest, stomach’, Badaic baramba ‘lungs’. Also note similarity
to Mal me-lembung ‘swell’ Together suggest < ?*ləmbəŋ
13
*baka is probably not an accurate etymon. *-k- is reflected only in Seko Padang, and there -a is strongly suspected
of being the trace of an echo vowel.
30
X
balesu: ‘rat, mouse’
SekP, SekT, Psn. Cf PKP *walesu'
*bànaŋ ‘thread’
SekP, SekT, Psn; Bdb bannang < probably Klp or Tab < PSS *bɨnnaŋ [PWMP *bənaŋ
(Blust ACD)]
*barako: ‘neck’ ?p
All; suggests < ?*ba-dakuq. [PIn ?*b(V)ruku ‘neck’ (Mills 626)] *u > *a unexplained,
probably a PSek innovation. Also, irregular *-∅ > *-:, but note PKP *waroko'. Cf PMP
*dukuq ‘bend over, stoop’ (Blust ACD), PBT *boroko (Mead ????), PSS *baroko ‘neck’,
Blt borokoʔ ‘throat’. PMP *dukuq should have yielded PKP *ruku'. If these etyma are
derived from PMP *dukuq, there must have been an innovation back there (PIn?) that
affected lots of Sulawesi microgroups. PMP ?*dukəq might better account for the vowel
lowering.
X
b/ma-baruʔ ‘chest’ (see *aruʔ)
SekT mabaru', Psn, Bdb babahu'. Unknown etymon; possibly related to PWMP *adəp
‘chest’ (Blust ACD), but how? Cf Bad bombaru', Ram bobari', PKP *bambara(')
*maN-bayaʔ ‘pay’
All [PMP *bayad (Zorc 1995)]
*beu: ‘lip’
SekP; Psn, Bdb kalambeu:, SekTH, SekTP kalambu:, SekTA kalambou14; suggests <
?*bequ. Cf Gor bihu ‘lip’, Bar wequ ‘raised inner lip (e.g. of a chest) on which the lid
fits’ (Mills 643); construction perhaps parallels Kul wiwihuu ‘lips’ from wiwi ‘rim’ + huu
(not reconstructed) ‘external mouth area’ (Mead 429)
*biiʔ ‘rear end, buttocks’
SekP, SekT, Bdb [PMP *birit (Blust ACD)] cf Mmj bui' < PMP *burit ; suggests? <
?*bərit
*maN-bika: ‘split (wood)’ cf bàka:
All [PMP *bikaq (Blust ACD)] This regular correspondence does not directly support PIn
*bi(ŋ)kas (Mills 1981); else > PSek **bikaʔ
*biluhak15 ‘hair (of head)’
Bdb; SekP and SekT bahulak (l-h metathesis and antepenultimate vowel lowering) Cf
PSS *belu(b)ak, Bad weluaq, Uma wulu wooq, PMP *buhək
*bintoeŋ ‘star’
SekT; Psn bintaŋ, *oe > a irregular, but also reflected that way in several PUS dialects
(Smith 163); Bdb bitue appears to be borrowed through a Kaili-Pamona language [PMP
14
Whether kalaN- is a live or frozen form, whether it reflects a root or functional prefix, and its semantic component
are uncertain.
15
*biluhak is presumably from ?*bulu baak ‘hair (of) head’. ?*bulu baak > ?*buluhaak > buluhak > biluhak. I have
no external evidence that the l-h metathesis happened pre-PSek. So perhaps this is one piece of evidence that another
node below PSek is common to Seko Padang and Seko Tengah.
31
*bituqən (Blust 1999)] pre-penultimate nasal insertion *∅ > *n probably happened preProto Seko; cf PSS *bintuin
*bìtiʔ ‘calf of leg’
All [PMP *bətis (Dempwolff 1938)] irregular if from 3-syllable PMP *bətiəs (Blust
ACD). Cf PSS *bɨttis
*bose ‘paddle’
Psn; SekP (pi-)bose , SekT pe-bose, Bdb bose: *e > e: unexplained [PMP *bəRsay ‘oar,
paddle’ (Dempwolff, in Zorc 1971 item 0916)]
*bua: ‘areca palm and nut’
All [PMP *buaq ‘fruit, areca palm and nut’ (Blust ACD)]
*buaya ‘crocodile’
SekP, SekT [PMP *buqaya (Blust 2002)]
*bulaʔ ‘ankle bone’ ?s
SekP also means ‘bud’ and ‘cockspur’; SekT bula'-bula' [PMP *bulat ‘stare with round
eyes’ (Blust ACD)] Phonological correspondence regular, but semantics uncertain,
sharing only the component of ‘round’ or perhaps ‘eyeball shaped’. Cf Mal bulat
‘spherical’. PMP *bulaŋ ‘artificial cockspur’ rejected as cognate; *ŋ > *ʔ cannot be
reconciled
*bulampa: ‘thigh’ ?s Cf lampa:
Psn, Bdb; SekP ‘long bones of legs or arms’ [PMP *lampaq ‘walk, go’ (Blust ACD), or
is this a homonym?] Probably from PSek *buu: ‘bone’ plus *mi-lampa: ‘long between
nodes’. *lampa: ‘bamboo water container’. Presumably named for the length
resemblance between a thigh and the container. Could it once have meant ‘walking
bone’? Another possible source: PMP buluq ‘slender bamboo’ plus (SekP) ampa:
‘internode’. Cf Tae bulo sanglampa ‘an internode of slender bamboo’ (Blust ACD).
*buloŋko: ‘spine, upper back’
SekP. Could be a low level innovation.
*miN-ka-buni (vi) ‘hide’ cf *maN-huni
SekTH, SekTP, Psn; SekTA me-ka-buni, Bdb miN-ka-la-buni, SekP miN-buni [PMP
*buni (Blust 1999)]
*buŋa ‘decorative plant’ cf *hùŋaŋ
SekP 1) ‘shrub sp, the flesh of which is eaten’, 2) ‘decorative shrub’, 3) (usually Redup)
‘decorative flower’; SekT (Redup in SekTH) ‘decorative flower’ [PMP *buŋa ‘flower’
(Blust 1999)]
*ma-busa: ‘white’
SekP, SekT, Psn [PMP *busa ‘foam’ (Blust ACD)] *a > *a: irregular. I posit PMP be
reconstructed ?*busaq ‘foam’. Anyway, that would place it closer to its double *bujaq.
Cf Klp, Tab, Bam, Rkg, Enr, Sad ma-busa ‘white’
*buto: ‘penis’
SekP, SekT, Bdb [PMP *butuq ‘testicles of animals’ (Blust ACD)]
32
*buu: ‘bone’
All; SekP also ‘core’ [PMP *buku ‘node, joint’ (Blust ACD)]
*dalaŋ ‘road, trail’
All [PMP *Zalan (Blust 1999)]
*daliʔ n ‘fly’
SekPH, Psn, Bdb [PMP *laləj (Dyen & McFarland 1970 item 352)] *l > *d irregular
dissimilation. Cf PPUS *daliʔ, KlpBH, Beh, Bad, Uma dali'. Probably inherited through
an intermediate node < ?*daliʔ
*ka-daluŋ ‘deep’ (see *ka-laduŋ)
X
daŋa ‘span’
Bdb [PMP *Zaŋan ‘hand span’ (Blust ACD)] cf PKP *(sa)-ndaŋa, Mmj daŋa
*daŋkaŋ ‘span’ ?d
Psn [PWMP *Zaŋkal ‘hand span’ (Blust ACD)] cf PPUS *daŋka(m/n), Sad daŋkan, PSS
*jaŋka(l?)
*dauŋ ‘leaf’
SekP, SekT; Psn dun-na, Bdb dau [PMP *dahun (Blust 1999)]
*mu-dauŋ ‘sew’
SekP, SekTA [PMP *ZaRum ‘needle’ (Blust 1999)]
*dio ‘you’
All [PMP ?*dikaw (find), else what is source of d?] Cf PMAT *dio
*i-dua ‘two’
SekP, SekT, Psn; Bdb dua [PMP *duha (Blust 1999)]
*dui ‘thorn’
Psn, Bdb [PMP *duRi (Zorc 1995)]
*m-eiʔ ‘cough’
SekP, SekT; frozen prefix [PWMP *ikəj (Blust ACD)]
*eno: ‘necklace’
SekP, Psn [PMP *hinuq ‘beads’ (Blust ACD)]
*maN-enuŋ ‘drink’
All [PWMP *inəm (Blust ACD), PMP *inum (Blust 1999)] The etymon *inəm fits best,
because it provides the conditioning environment for lowering *i > *e (see §4.2.1.4). But
also see §4.3.1.2 for *i > *e preceding PMP *u.
*m-esoŋ ‘sit’ ?s
SekP, SekT, Psn; Bdb misong [PMP *əsuŋ ‘rice mortar’ (Blust ACD)] 16 ma-ìsoŋ >
m-esoŋ. The relationship between the orientation of rice mortars and the sitting position is
16
PSek intransitive prefix *m- plus *ìsoŋ could have yielded either mesong or misong. The full form of the etymon
for the prefix here is uncertain. It could have been infix -um-. Or, it could have been prefix ma- or mi-. All are valid
intransitive affixes in Seko languages today. Perhaps the lowered vowel *e favors formation from a *ma- prefix.
33
logical, but should be investigated. Otherwise the PSek term derives from a separate
etymon. Cf PSS *isuŋ ‘sit’ (Mills 715)
*etik ‘marrow, brain’
SekP, SekT, Psn; SekTP itik [PMP *hutək (Blust ACD), probably with intermediate
node; *u > *e unexplained] Cf PSS *utɨk ‘brains’, Tab eti', Ram iti', Beh ati'
*maN-hàlaŋ ‘dry in sun, menjemur’ ?d
SekP. Etymon uncertain, expect < *bəlaŋ. Cf. PMP *bəkəlaj, Karo Batak pe-belang
‘spread out’
*hali ‘in balance’
SekP; SekP hali-ang ‘spouse, partner, one side (team) in a sports event, counterpart’,
SekP, SekTH, SekTP, Bdb maN-hali-i, SekTA me-hali, Psn miN-hali ‘reply, respond’,
SekP maN-hali-i also ‘come to someone’s aid, balance out a load (as on a horse), cut on
the opposite side of the first cut (as when felling a tree) [PMP *baliw ‘dual division,
moiety’ (Blust ACD)]
*hana-aŋ ‘shoulder’ ?s
All [PMP *baNaR ‘thorny vine, Smilax spp (for carrying firewood)’] cf PPUS
*aβa(n/ŋ)aN, Tab uhanaŋ ‘shoulder’. Also similar to PKP (East) *(a)waŋa, Bad uwaŋa,
but cannot reconcile PSek *-n- with PKP *-ŋ*hatu ‘stone’
SekPH, SekT, Psn, Bdb [PMP *batu (Blust 1999)]
*hauŋ ‘river’
SekT; SekP ‘stream’ [PWMP *bawaŋ ‘open expanse of land or water’ (Blust ACD)]
*heaʔ ‘rice between harvest and cooking’
All; SekP, SekT, and Psn extend the meaning to ‘rice before cooking’; SekP, SekT also,
‘year’ [PMP *bəRas (Blust ACD)]
*heŋi ‘night’
All [PMP *bəRŋi (Blust 1999)]
*hihi ‘edge, border’
SekP, SekT; Bdb sihi, unexpected assibilation *h- > s- (related to Mal sisi?) [PMP *bibiR
(Blust ACD)]
*hoa: ‘fruit’
All [PMP *buaq ‘fruit, areca palm and nut’ (Blust ACD)]
*hulaŋ ‘moon’
All [PMP *bulan (Blust 1999)]
*hulo: ‘slender bamboo of Schizostachyum genus’
SekT, Psn, Bdb; SekP ‘blowgun’, also, ‘the constellation Orion’s belt’ [PMP *buluq
(Blust ACD)]
*hulu ‘body hair, fur, feather’
All [PMP *bulu (Blust 1999, ACD)]
34
*hùŋaŋ ‘primordial fruit or its flower’ cf *buŋa
SekP, Psn, Bdb; SekP also, ‘fingers, toes’ [PMP dbl *buŋaŋ ‘flower’ (Blust ACD)] *u >
*ù irregular; suggests < ?*buŋa-aŋ
*hùŋiŋ ‘sand’
All [PMP *buhaŋin (Blust ACD)] cf PSS *buŋin
*ia (na)17 ‘if (it)’
SekP, SekT, Psn; Bdb ka naʔ 18 [PMP *ka (Blust ACD)] *∅ > *i unexplained
*ìloŋ ‘nose’
SekP, SekT, Bdb; Psn èloŋ19 (*ì > è unexplained) [PWMP *əjuŋ, dbl *ijuŋ (Blust ACD,
1999); *j > *l unexplained] suggests < ?*əluŋ or ?*idluŋ. Cf PSS ?*iloŋ, PPUS *illoŋ,
Dur, Klp, Rkg illoŋ, Bad ilo
*indaŋ ‘debt’
All [PWMP *hinzam ‘borrow, lend’ (Blust ACD)]
*isi ‘flesh’
All [PMP *isi (Blust ACD)] contrasts with PSS *ɨ(n?)si < PMP dbl *hesi (Blust ACD).
This contrasting pair affirms the reconstruction of doublets at the PMP level.
*issoŋ, dbl *ìsoŋ ‘rice mortar’
SekT, Bdb issong, Psn essong (irregular *i > e), SekP ìsong [PMP *iŋsuŋ dbl *əsuŋ
(Blust ACD)] cf Tab ingseng, Mmj insung
*maN-ita ‘see’
SekP, SekT, Bdb; Psn mu-ita [PMP *kita (Blust 1999)]
*kaa ‘elder sibling’ (possessive suffix required), *kaa: (vocative)
All [PMP *kaka (*kakaq vocative) (Blust 1979)]
*kabeʔ ‘left-handed’
SekP; SekTH, SekTP ‘left (side)’. Cf Bad, Beh kabeo' ‘left (side)’ Uma ngkabi' ‘lefthanded’ (rare word)
*kado:-kado: ‘throat’ ?d cf *kora:-kora:
Psn. Cf PPUS *kado-kado, Mmj kado kado, KlpBH kedo-kedo. PSek *-d- and many PUS
reflexes with -äd- (low front vowel) together suggest < ?*kanduq-kanduq, where the
nasal was subsequently lost. Note similarity to SekT kora:-kora:; probably related.
X
kaholoʔ ‘throat’ ?d
SekPL. possibly < unknown Kaili-Pamona language. Cf Uma tumolo ‘front of neck’ Ram
‘throat’, Bad tambolo ‘throat’, PKP *tambolo ‘front of neck’; but PKP *-∅ > PSek *-ʔ
irregular, same for frozen prefixes
17
*ia could be analyzed as derived from two morphemes *i-a. na is probably third person marker ‘3’, and the whole
sequence translated ‘if it be that’ or ‘in the event that’. The sequence a na should probably NOT be reanalyzed as àna
< *kəna, because in SekTH, SekTP, and Psn /n/ is doubled, which does not follow the pattern for PSek reflecting
*əC sequences.
18
The glottal in Bdb is unexplained. That, plus the retention of /k/, makes it a candidate for borrowing, from an
undetermined source.
19
quick vowel presumed; not written in data
35
*kahuŋ ‘cloud’ ?d
SekP [PAn dbl ?*ga(m)bun20 (Mills 699)] cf PMP *Rabun (Blust 1999), PSad *gawun
‘cloud’, PPUS *gaβun ‘fog’
*ka-iri ‘left side’
SekP, Bdb; SekTA keiri, Psn kehi; frozen prefix [PMP *ka-wiRi (Blust 1999)]
*kalae: ‘body’ cf *kale
SekP, Bdb. Cf PSS ?*kalawe / ?*kale; suggests < ?*kalaqi21
*kale ‘body’ cf *kalae:
SekT, Psn. Cf PSS ?*kalawe /?*kale; suggests < *?kalaqi
*kali-bambaŋ ‘butterfly’
SekT, Psn, Bdb; SekP kalu-bammang, *i > u irregular, but not uncommon to have
sporadic vowel shift in pre-penultimate position. Frozen prefix [PMP *kali-bambaŋ
(Blust 2001:20)]
*kalidi: ‘little finger’ ?s
SekP, SekTH, Psn, Bdb; SekP also, ‘twig’, SekTP ‘small one’ [PMP *kancil ‘animal’
(Dempwolff, in Zorc 1971 item 1737)] cf PKP *kanCili. Together suggest < ?*kanjiliq.
Presumed pathway to PSek: *kanjiliq > *kandiliq > *kadiliq > *kalidiq > *kalidi:. Note
similarity of PKP *kanCili to Mal kecil ‘small’ and kancil ‘mouse deer’(; but Mal lidi
‘palm leaf rib’ is also similar to SekP kalidi: ‘twig’.)
*kambaŋ ‘swell, swollen’
SekP, SekT; SekPH ma-kammang, SekT, Psn, Bdb ma-kambaŋ ‘thick, tebal’; Psn
kambang hacu (n) ‘boil’ [PMP *kambaŋ ‘bloom, well up, swelling’ (Dempwolff, in Zorc
1971 item 1729)]
*kambutuʔ ‘heel’
SekT, Psn. Cf PPUS *kambut(u/e)ʔ, Klp, Mms, Rkg kambutu'
*kànaʔ ‘complete, enough’
SekP; Psn ‘many’, SekT ‘perfect’ (Mills) [PAn *gənəp (Dempwolff, in Mills 702); also
PMP *kənaʔ ‘right, suitable’ (Blust ACD)] Either etymon would yield a regular reflex,
but the latter diverges semantically away from PSek. Nevertheless, both forms and
semantics are similar.
*kande: ‘eat’, also *ande:
SekP, SekT; Psn, Bdb m-ande:; suggests < ?*kandiq. Cf PSS *kande
X
kandiŋ ‘eyebrow’ cf *kìni
SekT, Psn; Bdb kandi ‘forehead’ (*ŋ > ∅ unexplained) [possibly < Mmj or Tab kandiŋ
‘forehead’ < PAn *kəniŋ ‘eyebrow’ (Mills 733)] irregular, expected > **kàniŋ or *kìniŋ.
Cf PSS *kɨnniŋ, PMAT *kandiŋ ‘forehead’
20
21
Mills based his proposed doublet on Dempwolff, which apparently contained uncertain *(m).
?*kalaqi could have followed independent pathways and yielded both PSek reflexes:
?*kalaqi > ?*kalaiq > *kalae:
?*kalaqi > ?*kalai > *kale
36
*kaŋkaŋ ‘forearm and hand’
All [PMP *gəmgəm ‘hold (in fist)’ (Blust 1999)]
*kanoni: ‘heel’ ?d
SekP
X
maN-kàru: ‘scratch, scrape’
SekP, SekTH. [< Ram mangkokaru' < PKP *karu' < PMP *kadus ‘scratch, rub’
(Dempwolff, in Zorc 1971 item 1681)] maN-ka-karu' > maN-ka-aru' > maN-kà-ru' >
maN-kàru:
*karuruŋ ‘skull’
SekP, SekT, Psn. Cf Klp karurung, Rkg karuru, PPUS *karorak ‘outer shell (skull,
coconut shell)’
X
katinimpu: ‘thumb’
SekPL; SekPH katinempu:, Psn kinampu [< ? < PMP *t-impu ‘grandparent’ (Blust
ACD)] *∅ > *: irregular. Uncertain inheritance path, probably indirect. Cf Ram
kototumpu, Bad katumpu
*mi-katiŋ ‘itch’
SekP, SekTA, Psn; Bdb ma-katiŋ [PMP *katəl (Dempwolff, in Zorc 1971 item 1761),
PWMP *gatəl (Blust ACD)] Languages with -ŋ cognates: Tab, Mmj. Languages with -ʔ
cognates: Klp, Sad, Dur, Bam. *ə > *i here may be diagnostic for classification; cf Mills
(p 728) and Martens (p 199)
*maN-kau: ‘beckon’
SekTH; SekP ‘lead, entice?’ [PMP *k-aku ‘to me’ (Blust ACD)]
*kayu ‘wood’
All; SekPL, SekT also ‘tree’, SekPH some speakers ‘firewood’ [PMP *kahiw (Blust
1999)]
*keboʔ ‘back’
All [PMP ?*bu(ŋ)ku(jt)22 (Mills 660)] b-k metathesis assumed, else not related to PMP
etymon; *u > *e unexplained, expected < *ə somewhere. Cf PSS boko(t?) and PSS
?*keboC (Mills 888), PMAT *keboʔ
*maN-kekiʔ ‘bite’
SekP, SekT [PMP *kitkit (Blust 1973)]
*kìni ‘nail, claw, hoof’ cf Xkandiŋ ?s
All; SekP kìni mata ‘eyebrow’ [PAn *kəniŋ23 ‘eyebrow’ (Mills 733); *ŋ > *∅24] cf PMP
*kinit ‘pinch, nip, pluck’ (Blust ACD)
22
Mills suggested his proto-form could also reflect PAn *buku ‘bone’.
PAn *kəniŋ ‘eyebrow’ apparently developed along two pathways. *kəniŋ > *kìniŋ > *kìni is the more regular for
PSek, but the meaning ‘nail, claw, hoof’ suggests a more generic PAn gloss, like ‘hard, protective
covering/protrusion’. The other pathway, probably through PSS, doubled the nasal, which then became -nd- in
languages where geminate nasals are disallowed root medially. *kəniŋ > *kɨnniŋ > kandiŋ.
24
Loss of *-ŋ could be explained by back formation. The root normally occurs with a possessive suffix, so it might
get reinterpreted as a nasal ligature (see §2.4). *kìniŋ-na > *kìni-nna ‘its hoof’ (both pronounced identically
/kìnin:a/)
23
37
*kokoʔ ‘leg, foot’
All [PMP *kukud ‘shank or hoof of animals’ (Blust ACD)] irregular *u > *o, suggests <
?*kudkud or ?*kukəd
X
kolla ‘sugar’
SekP, SekT, Psn; Bdb golla. [< Mal gula ‘sugar’ < Skt guḍa, gula ‘molasses’ (Mills
899)]
X
kolloŋ ‘throat’ ?d
SekPH. < (likely) Rkg kollong ‘neck’ < PSS *kɨlloŋ ‘neck’. Related to? PMP *kəluŋ
‘coil, curl’ (Blust 1970, in Zorc 1971 item 1834)
*kompoŋ ‘belly’
Bdb; Psn pompong, *k > p irregular. See Mills 733 for similar irregularities. [PMP
*kəmpuŋ ‘stomach of animals’ (Zorc 1995)]
*kona: ‘name’
All [PIn ?*guna(q) (Mills 1981 item 17)] *u > *o irregular; but o in all reflexes
*kora:-kora: ‘throat’ ?d cf *kado:-kado:
SekT. Cf Rkg gora-gora. Together suggest < ?*gudəq-gudəq, else indirectly inherited.
*koriŋ ‘cooking pot’
All [PWMP dbl *kudən (Blust ACD)]
*k-uanaŋ ‘right side’
SekTH, SekTP, Psn; SekTA koanaŋ, SekP konaŋ, Bdb kanaŋ [PMP *ka-wanan (Blust
1999)]
*kuliʔ ‘skin’
All [PMP *kulit (Blust 1999)]
*kuma: ‘sheath’
All [?*gumaq25 < PMP *Rumaq ‘house’ (Blust 1999)]
*kuntuʔ ‘knee’
All. Cf PPUS *guntuʔ ‘knee’, PSad *guntu(t?) (Mills 882), Bad kuntuʔ, Ram, Uma kotuʔ
*kutu ‘head louse’
SekP, SekTA, Psn, Bdb [PMP *kutu (Blust 1999)]
*ka-laduŋ ‘deep’ also, *ka-daluŋ
SekP, SekT, Psn; Bdb ma-ka-laduŋ, SekT also, ka-daluŋ. Frozen prefix ka- [PMP *idaləm ‘in, inside, deep’ (Blust 1999, ACD note)] d-l metathesis. Variance between *ə >
*u in *ka-laduŋ and *ə > *i in *i-laliŋ unexplained
25
Although ?*gumaq ‘sheath, housing’ has not, to my knowledge, been reconstructed at an intermediate level, PSek
is clearly inherited from this form. The phonological correspondences are regular, but not so with *Rumaq. There
are outside witnesses from Mmj, Mdr and Bar guma ‘sheath’, and PKP *guma(') ‘k.o. sword or sheath’ (semantic
shift?). PPUS constructs banua piso and Sad. banua la'bo' (Lit., ‘machete housing’) for ‘sheath’, from banua
‘house’, corresponding semantically to PMP *Rumaq ‘house’ > ?*gumaq.
38
*ma-lai ‘leave for home’
All; SekP also, ‘go to a more suitable place’ [PMP *laRiw ‘run, run away, flee, escape’
(Blust ACD)]
*laia ‘ginger’
SekP, SekTH, SekTP, Psn, Bdb; SekTA leia [PMP *laqia (Blust pers. com. to Mead)]
*mi-lalaŋ ‘spicy hot’
SekP; Psn, Bdb ‘ginger’ [PMP *laja ‘burn, smart (a wound), be hot (spices)’
(Dempwolff, in Zorc 1971 item 2043)] cf Sad, Dur ma-lala. *-∅ > *-ŋ unexplained;
possibly a loan inherited through Bug ladang < Mal lada ‘pepper’ (Mills 908)
*laliaŋ ‘three-stone fireplace’
All [PMP *dalikan (Blust 1999)] *d > *l assimilation regular. Tld lalikang; borrowed?
*i-laliŋ ‘inside’
SekP, SekT; Psn i-laiŋ, *l > ∅ unexplained, but several neighboring languages have
similarly lost one of the l’s [PMP *i-daləm ‘in, inside, deep’ (Blust 1999, ACD note)]
*lambeʔ ‘back of the knee, popliteal space’
SekP, Psn, Bdb; SekT lambeʔ-lambeʔ. Cf PPUS *lambeʔ ‘calf of leg’, KlpBH lambe'
*lampa: ‘bamboo water container’ ?s
SekT; SekP ‘stack of food containers, rantang’ [PMP *lampaq ‘walk, go’ (Blust ACD)]
cf PPUS *lampa, Rkg, Klp lampa
*laŋiʔ ‘sky’
All [PMP *laŋit (Blust 1999)]
*laŋka:-laŋka: ‘spider’ ?s
All [PMP *laŋkaq ‘step, stride’ (Dempwolff, in Blust ACD note)]
*laŋke: ‘shackle’
SekP
*lao ‘go, walk’ (variously affixed)
All [PMP *lakaw (Blust 1999)] The first k in SekTP mellaok-king appears to reflect a
metathesized k from proto-form *lako. I don’t know enough about SekTP
morphophonemics to say for sure. Badaic l-um-ao
*ma-lea: ‘red’26
All. Suggests < ?*ma-ləRaq. Cf PSad ?*lea, PKP *ma-lei, PMP *ma-iRaq. Related to
PBT *leaº ‘flame’?
*lèsok ‘hobble’
SekTP lèsok-kang ‘ankle (not the joint)’, SekP mimpa`lèsoko ‘scoot oneself forward or
around on the floor’, Psn lèso'ang ‘ankle joint’; quickness uncertain in SekTP and Psn. Cf
PKP *leso' ‘?ankle/wrist joint’, Lindu lesok-a (frozen suffix) ‘joint’
26
I think I have heard that, in the practice (archaic in Seko Padang) of making red dye, a plant called lea:, which I
have yet to see or identify, is cooked. But this needs to be verified.
39
*lila: ‘tongue’
All [PMP *dilaq (Blust 1999); Sirk claims (p 58) *lilaq is pre-PSS, the same would also
hold for PSek]
*lima ‘five’
All [PMP *lima ‘hand, five’ (Blust ACD note)]
*lindo: ‘forehead’
SekP, SekT [PMP *li(n)juq, PIn ?*li(n)juq ‘face, forehead’ (Mills 330-1, 765)]
*ti-loa: ‘vomit’
All; SekP maN-loa: ‘spit out’ [PWMP *luaq ‘expel from mouth’ (Dyen 1953), dbl
‘empty out’ (Blust ACD)] *u > *o unexplained, but same pattern as *hoa:
*ma-lotoŋ ‘black’ ?d
SekP [PMP *lutuŋ ‘monkey’ (Blust 2002)]
*ba-luaʔ ‘wide, spacious’
SekP, Psn, Bdb. Frozen prefix [PMP *ma-luqas (Zorc 1995)]
*màmiʔ ‘delicious’
SekP. Frozen prefix [PMP *ma-əmis ‘sweet (taste)’ (Blust ACD note)] possibly related to
PSek *maN-tàmi' ‘taste’
*manok, dbl *manuk ‘chicken’
SekP `manoko, SekT manok, Bdb mano'; Psn manu' [PMP *manuk ‘hen, bird, animal’
(Dempwolff, in Zorc 1971 item 2438), ‘bird’ (Blust 1999)] Tld mano'. Irregular *u > *o,
perhaps < ?*manək. Cf Tab mane', Bam mäne'
*mata ‘eye’
All [PMP *mata (Blust 1999)]
*mate ‘die, dead’
All [PMP *m-atay (Blust 1999)]
*muane ‘male’
SekP, SekTA; SekTH, SekTP mane, Psn manane, Bdb muniane; the latter two based on
reduplicated *muane-ane [PMP *ma-Ruqanay (Blust 1999)]
*muŋku ‘mountain’ ?d
SekPH. Cf PSS ?*buŋku (Mills 659), Klw (Mills), MorB, Pad tamuŋku, Bar muŋku (ritual
language), Top muŋgu ‘mountain’; Daa buŋgu ‘summit’. Cognates with m- are in Central
Sulawesi languages, so possibly a loan.
*mi-na-naha ‘breathe’
All; Tld *mi-naha [PMP *ñawa ‘breath’ (Dahl 1981b)]
*mu-ma-nau ‘steal’ ?d ?p
SekPH; second morpheme break frozen [PMP dbl *nakaw (Blust ACD)] irregular *-aw >
*-u, expected mu-ma-nao. Perhaps *nakaw > *nako > *nao > *nau. Cf Ram, Beh manako, retaining medial *k; Uma ma-nako
40
*nau: ‘go downstream’
SekP, SekT. Cf PPh *naqu(gR) ‘descend’ (Charles 1973), PKP *na'u ‘descend, go
downhill’
*maN-oa ‘say’
SekP, SekTH, SekTP, Psn. Cf PSS *kua
*oaʔ ‘root, vine’ cf *uaʔ
All; one SekPH witness boa', b unexplained [PWMP *uRat ‘blood vessel, nerve, tendon,
root’ (Blust ACD); alternatively PMP *Rawəj, dbl *waRəj ‘vine, aerial root’ (Zorc 1995)]
cf PSS *urat
*oriŋ ‘charcoal’
All [PMP *qujiŋ (Blust ACD)] *u > *o irregular vowel lowering (< ?*qujəŋ would be
better fit, due to vowel lowering in *ə environment)
*osok ‘rib, side’
All [PMP *Rusuk ‘rib’ (Dempwolff 1938)] cf PSS *rusuk (Mills 299), PPUS *usuk
*otoʔ ‘little crawly parasite’
SekP ‘maggot’; SekTH, SekTP ‘head louse’; Psn, SekTH ‘caterpillar’; Psn oto' malea:
‘red chicken mite’ [PIn ?*rutus ‘worm sp (marine), harmful to boats’ (Mills 694)]
*paa: ‘thigh’, *paa (Possessive form)27
SekP; SekTA paang [PMP *paqa (Blust ACD)]
*pàduŋ ‘heel’ ?d
Bdb. Cf PKP ?*pad(u/o)'
*ma-paiʔ ‘bitter’
Psn, SekTH; SekP mi-pai', SekTA, SekTP me-pai', Bdb pai' [PMP *paqit (Blust ACD
note)]
*palaʔ ‘palm of hand, sole of foot’
All [PMP *palaj (Blust ACD)]
*palo ‘thigh’ ?d
SekTH, SekTP. Low level innovation?
*maN-pana: ‘shoot (with bow and arrow)’
SekP, SekTA; SekTP, Psn, Bdb maN-pana; SekP pana: 28 ‘bow’ [PMP *panaq ‘shoot’
(Blust 1999)]
*ma-panaʔ ‘sick, in pain’
Bdb; SekP mi-pana', SekTH, SekTP me-pana' [PMP *ma-panas ‘warm, hot’ (Blust
1999)]
27
Although SekP paa: reflects metathesis of *-qa, the possessive construction paa-nna ‘his thigh’ retains a nasal
ligature, which can only attach to the unmetathesized form, so we reconstruct a second PSek etymon without vowel
length. Also, one witness claimed that the reflex paa was used by the older generation. SekTA paang could be a
back formation from the Possessive form. On the other hand, cf Mmj, Tab paang, which appears to be a metathesis
of PPUS *apa(m/n)
28
In Panasuan pana was claimed to be a new, introduced item. In Seko Padang, it is only known as a child’s toy. No
term for it in Seko Tengah - Hoyane.
41
*paniʔ ‘wing’
All [PMP *panij (Blust ACD)]
*pape: ‘cheek’
Psn, Bdb; suggests < ?*papiq. Cf PPUS *pap(o/e)
*pare ‘rice plant’ ?d
Bdb; Psn ‘year’ [PMP *pajay Blust ACD)] regular, but possible loan from Klp
*pira ‘how many?’
SekP, SekT, Psn [PMP *pija ‘how much?, how many?’ (Blust ACD note)]
*piso ‘machete’ ?d
SekP [PAn *pisaw (Dempwolff, in Zorc 1971 item 2764)]
*pitu ‘seven’
All [PMP *pitú (Zorc 1995)]
*pompo: ‘blunt, dull (point)’
All [PMP *pul (Zorc 1995)] presumably *pulpul > *pumpuq > *pompo:. Cf PSS
*pompoŋ (Mills 891)
*puduk ‘mouth’ ?d
Psn. Correspondence regular, but possibly borrowed. Cf Tld pudu', Rkg puduk, Botteng
pudu' ‘nose’, Mms puduk ‘lip’; suggests < ?*puduk ‘snout, moncong’. Note PPUS
*puduN ‘nose’, SekP paropok ‘snout’
*pulo: ‘ten’
All [PMP *púluq (Zorc 1995)]
*maN-puŋu ‘bind, shackle’
SekP; Psn ka-pungu-ang ‘ankle (not the joint)’ [PMP *puŋu ‘bunch, cluster’ (Blust
ACD)]
*pùno: ‘full’
SekP, Psn, Bdb [PMP *pənuq (ACD note, or PAn *pənúq Dyen & McFarland 1970 item
468)]
*punti ‘banana’
SekPL, SekT, Psn, Bdb [PMP *punti (Dempwolff, in Zorc 1971 item 2822)]
*pùru ‘gall (bladder), bile’
SekP; SekTP pa'du, Bdb pàdu, Psn pàru29 [PMP *qapəju (Blust ACD)] *qa- > *∅*puso: ‘heart’
SekP [PMP *pusuq (Dyen 1953)]
*raa: ‘blood’
All [PMP *daRaq (Blust 1999)]
29
The SekTP form pa'du is certainly borrowed, possibly from Rkg, known by the root medial glottal. The Psn and
Bdb forms are irregular, and are possibly Seko-ized versions -- pa'du > pàdu. An alternative analysis, which I
consider less likely, is to reconstruct PSek *pàdu. Under that scenario the change to SekP pùru would have occurred
post-PSek. Although such a progression is regular, the timing is late. It would be the only intervocalic PMP *j to be
reconstructed PSek *d, and the only instance of PSek *d > SekP r.
42
*rakiʔ ‘raft’
All [PMP *dakit (Blust ACD)]
*raŋkaʔ ‘fingers, toes’
SekT, Psn, Bdb; SekP ‘crab’. Uncertain etymon. If directly inherited, expect < ?*daŋkaC.
Cf PPUS *reŋkeʔ / raŋkaʔ ‘finger’, Rkg raŋka' (glottal presumed, not written), Klp
rakka', ‘fingers, toes’, PMP *kaRakap ‘crab sp.’ Possibly indirectly inherited from
PWMP *Zaŋkal ‘handspan’ (Blust ACD), cf PSek *daŋkaŋ ‘span’
*ka-rao: ‘far’
All; frozen prefix [PMP *ma-Zauq (Blust 1999)] cf Nap karao, Beh, Bad karao',
perfectly corresponding to PSek but unlike any other Kaili-Pamona languages.
*rasaŋ ‘house’
SekP [PIn *dasan (Mills 1981)] cf PKP *rasa ‘hut’
*rea: ‘sword grass, Imperata cylindrica’
SekP, SekT; Psn ha-bahi [PMP *Riaq (Zorc 1995)] *R > *r irregular, expected <
?*dəRaq; probably inherited through an intermediate node. Cf PPUS *rea, Mak rea, Tam
reaʔ
*rindiŋ ‘wall’
SekP, SekTH, SekTP [PMP *diŋdiŋ (Dempwolff, in Zorc 1971 item 1420)]
*ka-roa-a: ‘eight’ cf *dua
SekT, Psn, Bdb; SekP karo'a:. Frozen affixes [PMP *duha ‘two’ (Blust 1999)] *u > *o
irregular for SekP, more typical of SekT
*rombe ‘hair (of head)’
Psn; alternate hambe, younger generation. Cf PSS *(d)ombe ‘fringe’, Mak rombeng
‘long-haired’ (Mills), Mal rumbai ‘tuft, tassel’, Kul rembe ‘body hair’ (Mead 491)
X
rupa ‘face’
Psn, Bdb. [< Skt]
*sàbu ‘thousand’
SekP, SekT, Psn; Bdb sàbu: *u > u: unexplained [PMP *sa-Ribu (Dahl 1981a)] *sa-Ribu
> *sa-ibu > *səbu > *sàbu
*sadaŋ ‘chin, jaw’
SekPH, SekT, Bdb; SekPH also ‘gill’; Psn sadaŋ, expected saraŋ30. Cf PSS *sadaŋ
*saha ‘python’
SekP; Bdb ‘snake’ [PMP *sawa (Blust ACD note, 2002)] possibly also applies to SekT
and Psn, limited data
*sai ‘illness’
SekP, Psn [PMP *ma-sakit ‘sick, painful’ (Blust 1999)] *t > *∅ irregular, but PSS *saki
also shows unexplained loss of final consonant; same with Badaic
30
Mills also noted irregular retention of /d/ in Makassarese (p 811)
43
*sali ‘floor’
SekP, SekTH [PMP *saləR (Dempwolff, in Zorc 1971 item 2932)]
*sape: ‘chin, jaw’ (possibly *sàpe:) ?s
SekPL [PAn ?*səpiq (Mills 824, modifying Zorc *sipiq ‘section, part, strip off’)] Cf PSS
*sɨppe ‘break, separate’
*sia ‘salt’
SekPH, SekT, Psn, Bdb [PMP *qasiRa (Blust 1999)]
*siku ‘elbow’
Psn, Bdb; SekP, SekT siu [PMP *siku (Blust 1999)] Beh, Nap hiu31
*sila ‘face’ ?s
SekP, SekT; Psn ‘forehead’ [PMP *sila ‘outpouring of light’ (Blust ACD)]
*sise: ‘tooth’ ?d
SekP; suggests < ?*sisiq [PWMP *sisig ‘pick or clean the teeth’ (Blust ACD)] irregular
*-g > *-:
*maN-sisiʔ ‘suck on’
SekP [PMP *səpsəp ‘suck’ (Blust 1999)]
*sua: ‘mouth’
SekPL, SekT, Bdb; SekPH ‘voice, word’; SekTA soa:. Cf PMAT *sua, Mal suara
‘voice’ < Skt svara; suggests < ?*suaRa or ?*suRaq or ?*suqa. But note Ledo sumba
‘mouth’, PMP *suba/supaq ‘spittle’
*s-um-ule: ‘return home’ ?s cf maN-ule:-kiŋ
SekTH, SekTP, Psn; SekTA s-um-ule, SekP ‘return’, SekP maN-sule-i: ‘return to look for
something’ [PMP *suliq ‘repay, requite’ (Charles 1973)] probably related to PMP
*um-uliq ‘return home’ (Blust ACD), but *∅- > *s- is problematic. Cf PSS *ole (Mills
879)
X
sumpi: ‘tongs’
SekP, SekTH. < unknown source sumpiq < PWMP *su(m)pit (Blust ACD) PMP *t >
PSek *: irregular; expected > **sumpiʔ. Cf PKP *supi'
*sura ‘thorn’
SekP, SekT; SekP also ‘short spear or dagger made from e.g. bamboo, ranjau’ [PMP
*suja ‘pitfall spike’ (Zorc 1995)]
*sùruʔ ‘sip, suck, slurp, inhale’
All [PWMP *sədut ‘sip, suck’ (Blust ACD)]
*susu ‘breast’
All [PMP *susu (Blust 1999)]
X
tabaro: ‘sago palm’
SekP, SekT; Psn cubaho:, cf Ram tubaro'. [< unknown Celebic language tabaro']
31
The PSek reconstruction reflects later loss of *k, with subsequent borrowing into Behoa and Napu. An analysis
reflecting earlier loss of *k might also be possible, or a reconstruction of doublets.
44
Irregular *r > *r. If directly inherited from PMP, expect < ?*tabaduq, else if etymon =
?*tabaruq, expect *r > *∅, > **tabao:. Cf PKP *tabaro'.32 Cf PMun *tabaro ‘sago
flour’. Mills (p 912) also claimed this as a loan for South Sulawesi languages
*mi-taha ‘laugh’
SekTP, Psn, Bdb; SekP, SekTA, SekTH ma-taha [PMP *tawa (Blust 1999)]
*tai: ‘guts, feces’
All [PMP *tinaqi (Blust 1999), *t-in-aqi (Blust ACD note)] metathesis *-qi > *-iq > *-i:
*tako: ‘mouth’ ?d
SekPH; suggests < ?*takuq. Local innovation? No known correspondences
*tàlaŋ ‘bamboo sp’
SekP, SekT [PWMP *təlaŋ (Blust ACD)]
*taliŋa ‘ear’
All [PMP *taliŋa (Blust 1999)]
*i-tàlu ‘three’
SekP, SekT, Psn; Bdb tallu < Klp or Mmj [PMP *təlu (Blust 1999)]
*talubambaʔ ‘snake’ ?d
SekPH. Frozen morpheme break possible *talu-bambaʔ or *ta-lubambaʔ or both.
Possible local innovation; find outside cognate or delete
*tambusu ‘cheek’ ?d ?s
SekP, SekT [PWMP *busuR ‘satisfied from having eaten enough; satiated’ (Blust ACD)]
Frozen prefix taN- presumed, but uncertain meaning. Could be lower level lexical
innovation
*tanduk ‘horn’
All [PMP *tanduk (Dempwolff, in Zorc 1971 item 3417)]
*pa-taruŋ ‘sharp’
SekP [PMP *ma-taZəm (Blust 1999)] Nap, Beh taru, Bad tumaru. Cf PKP *ma-taja. The
Badaic are similar to PSek, not to PKP
X
tarutu ‘chest’ ?d
Psn. Possible loan from KlpKr. Cf KlpKr tarutu, Top tanutu
*tasik ‘sea’
SekP, SekT; Psn sasi' (probable loan, Mdr sasi' (Mills), KlpBH sasik) [PMP *tasik (Blust
1999)]
*tèdeʔ ‘stand’
All [PMP *tənZəg ‘upright, erect’, or *ti(n)Zəg ‘stand erect’ (Blust ACD)] *nZ > *nd >
*d
32
Martens (1990:165) wrote, “Sago was virtually unknown in the Uma mountains until 50 years ago. The term is
therefore probably a borrowing in some KP languages and not a native word.”
45
*teme: ‘urine’
SekPH, SekTA, SekTH, Psn, Bdb [PMP *kəmiq ‘urinate’ (Blust ACD)] *ə > *e irregular,
expected *è; *k > *t irregular.33 Cf PPUS, PSS *teme, *tene
*maN-tibe ‘throw away’
All [PIn *ti(m)baR (Mills 1981:71)] irregular *-aR > *e, expected < ?*ti(m)bay or
?*ti(m)bəR. Cf PSS *ti(m)be
*timbiʔ ‘span’ ?d
SekP, SekT. Cf Ram timbiʔ. Appears to be lower level innovation or loan from Rampi,
which in either case should not be reconstructed. But, hold until certain of no other
cognates.
*toaʔ ‘base of tree’
SekP, Bdb [PMP *tuqəD ‘tree stump, stubble’ (Blust ACD)]
X
tobok ‘kris’, XmaN-tobok ‘stab’ cf *maN-tòbok
SekPH, SekT [< Sad? tobok (Mills 863) < PMP *təbək ‘pierce, stab’ (Blust ACD)] If
directly inherited from PMP, *ə > *o irregular; expected > *ò, therefore considered loan.
*ma-tua: ‘mature, old (living things)’
All; SekP also, ‘hard’ [PMP *ma-tuqah (Blust 1999)] metathesis regular *qa > *aq > *a:
*tuho: ‘live, grow’
All [PMP *tu(m)buq ‘grow’ (Blust 1999), *tubuq ‘sprout’ (Blust ACD note)]
*tùhu ‘sugarcane’
SekP, SekTA [PMP *təbuh (Blust pers. com. with Mead)]
*tùlu ‘egg’
SekP [PMP *qatəluR (Blust 1999)]
*maN-tunu ‘burn’
Bdb; SekP ‘roast’ [PMP *tunu (Blust 1999)]
*maN-turo: ‘point’
SekP, SekT [PMP *tuZuq (Dyen, in Zorc 1971 item 3662)] *tuZuk (Dyen & McFarland
1970 item 681) rejected because *-uk > *-o: is irregular
*ma-turu ‘lie down’
SekP, SekTH, SekTP; SekTA ‘go to bed’ [PMP *tuduR ‘sleep’ (Blust 1999)]
*t-um-utu ‘pound rice’
SekP, Psn; Bdb t-im-utu, SekP maN-tutu ‘pound’ [PMP *tutu ‘pound’ (Blust 1999)]
*uaʔ ‘blood vessel, nerve, tendon’ cf *oaʔ
SekT; SekPH oaʔ , SekPL uraʔ , Psn, Bdb uhaʔ 34 [PWMP *uRat ‘blood vessel, nerve,
33
Mead (pers. com.) suggests possible early loss of k-, then addition of a verbal prefix having t. Cf MorB te'eme
‘urinate’, PBT *eme ‘urine’. While this analysis is possible, I am doubtful, because, in addition incorporating an
unknown verbal prefix, the form would have to be reanalyzed from a verb back to a noun. *kəmiq > *əmiq >
*t-emeq (v) > *temeq (n). The verbal infix -um- is employed in modern day SekP, t-um-eme: ‘urinate’.
34
The SekPL, Psn and Bdb reflexes are loans, probably from neighboring Rkg and Klp. SekT maintains a distinction
between uaʔ ‘vein’ and oaʔ ‘root’, whereas SekPH does not. It is uncertain how that came about.
46
tendon, root’ (Blust ACD)] cf PSS *urat ‘vein, tendon; root’, PPUS *uraʔ, PKP *ua',
‘vein (blood)’
*uhai ‘water, river, drink’
All [PMP *wahiR ‘water (fresh)’ (Blust 1999)] *w- > *uh*uhe ‘rattan’
SekP, SekT [PMP *quay,35 or *buay ‘rattan fruit (?)’ (Blust ACD)] *u- > *uh- irregular
(but cf *uhai and *uhati); expected < ?*quway;36 else, b-u metathesis, *buay > *bue >
*ube > *uhe. Cf Bad, Beh uwe
*maN-ule:-kiŋ ‘return something, return home’ cf *s-um-ule: ‘return home’
SekP; SekTA maN-ule-kiŋ, SekP ba-ule:-ule: ‘go to and fro’ [PMP *maN-uliq ‘return
something; restore’, *uliq-an ‘go home’, *uliq uliq ‘repeatedly’ (Blust ACD)] PSS *ole
‘again, return’ (Mills 879).
*ulo ‘tear, air mata’ ?s
Psn archaic/obsolete; SekP ulo: (*o > o: unexplained, suggests < **uluq) [PAn *ulaw
‘confusion, perplexity’ (Blust ACD)]
*ùnuŋ ‘six’
All [PMP *ənəm (Blust ACD)]
*ùpaʔ ‘four’
SekP, SekT, Bdb; Psn òpa', sporadic vowel lowering [PMP *əpat (Blust 1999)]
X
uraʔ ‘vein (blood)’
SekPL, Psn, Bdb; SekP also, ‘lines in one’s palm’. [< Rkg uraʔ or Klp < PPUS *uraʔ,
PSS *urat ‘vein, tendon; root’ < PWMP *uRat ‘blood vessel, nerve, tendon, root’ (Blust
ACD)]
*uraŋ ‘rain’
SekP, SekT, Psn [PMP *quZan (Blust 1999)]
*uruŋ ‘fish’ ?d
SekP. Cf Uma uru ‘fish’37
7.3
PSek wordlist reconstructed based on Seko Padang only
*maN-ala ‘take, get, fetch’
[PMP *ala (Blust ACD)]
35
Mills (1975:876) offered PSS ?*we as an alternate reconstruction. That form would serve as a regular etymon
intermediate between PMP and PSek.
36
Charles (1973) reconstructed PPh *quway and PAn *quwey ‘rattan’, but -w- forms were rejected by later
researchers for those levels (where did I read that?), so contrary to first impression, PSek *uh must not reflect PMP
*uw.
37
?*uru(N), if there was such a proto-form, meant ‘snakehead fish’. I don’t know if that fish was introduced or is
native. Ledo bau uru ‘snakehead fish’, Moma, Lindu uru ‘snakehead fish’, Uma uru mpu'u ‘snakehead fish’ (lit.,
‘real/true fish’) (Martens, pers com), SekP urung maroa' ‘snakehead fish’ (lit., ‘common fish’, but could also mean
‘original fish, ikan asli’)
47
*aloʔ ‘knobbed hornbill’
[PMP dbl *alaw (Blust ACD)] *aw > *oʔ irregular, probably a spontaneous innovation;
also unexplained in Uma
*m-ànaʔ ‘be quiet, calm’ ?p
Prefix presumed, frozen; cf maN-àna' ‘make straight’, mu-ling-àna' (second morpheme
break presumed) ‘lie at full length’. [PMP *qənəŋ ‘quiet, still, at rest’ (Blust ACD)] *ŋ >
*ʔ irregular, but cf CEMP *qənəp ‘lie down to sleep’ where p > ʔ is regular
*anitu ‘ghost, spirit’
[PMP *qanitu (Blust ACD)]
*ara ‘be, exist, have, wealthy’
[PMP *wada (Blust ACD)]
*maN-àruʔ ‘scrape, shave (as rattan), whittle to a point’
[PWMP *aRud (Blust ACD)] irregular *R > *r and *a > *à; expected < ?*əjud
*maN-asa: ‘file teeth’
[PMP *hasaq ‘whet, sharpen’ (Blust ACD)]
*maN-baha ‘carry, bring, lead’
[PMP *baba ‘carry on back’ (Blust ACD)]
*bàka: ‘crevice’
distinct from *bika: [PMP *bəkaq ‘split, crack open’ (Blust ACD)]
*maN-bala: ‘shape with a machete by cutting away slivers with the grain of wood’
[PWMP *bala (Blust ACD)] *a > *a: irregular. Suggests PWMP etymon be revised to
?*balaq
*balèbeʔ ‘cockroach’
*bàtik ‘speckled, blotched (as markings on a chicken or snake)’
[PMP *bətik ‘tattoo’ (Blust ACD)] Uma woti' ‘etch, decorate’. Tentative PKP *woti'
‘etch, decorate’ or ‘speckled’
*bebe ‘mute, stupid’
[PMP *bəlbəl ‘plugged, stupid’ (Blust ACD)] irregular?38 *-l > *-∅, expected? <
?*bəRbəR. PKP *bobo(') ‘mute, stupid’. But Behoa has bebe
*benuʔ ‘husk’
[PWMP *bənut ‘coconut husk’ (Blust ACD)] irregular *ə > *e, an irregularity common
throughout Sulawesi (Mead, pers. com.)
*maN-bobok ‘dig horizontally’
[PWMP *maN-bəkbək ‘pound into fine particles’39 (Blust ACD)]
38
No actual pattern has been established for sound change from etyma ending in l. Only two other correspondences
are known to me: PMP *katəl ‘itch’ > *PSek *mi-katiŋ, where *l > *ŋ, and PMP *pulpul ‘blunt, dull’ > PSek
*pompo:, where *l > *q > *:.
39
Whereas most reflexes assign a meaning related to the pulverizing done by weevils, it appears that the Seko term
is related to the work of boring. Blust calls PMP *bukbuk and *bəkbək ‘weevil (dust)’ doublets (ACD note), but
does not reconstruct *maN-bukbuk.
48
*bubuk ‘frass, leavings’
[PMP *bukbuk ‘weevil, dust produced by weevil’ (Blust ACD)]
*bulai ‘co-parents-in-law’
Cf PPh *balaqih (Blust ACD). Note that *q was lost and did not metathesize with *i to
yield PSek *-i:. Also, *a > *u irregular. Nevertheless, this reflex should encourage
reconstructing the etymon at a higher level.
*maN-buno: ‘take heads’
[PMP *bunuq ‘kill’ (Blust 1999)]
X
damaʔ ‘resin’ also, ‘tree that produces (copal) resin, Agathis sp’
[< Mal damar < PMP *damaR ‘resin’ (Zorc 1995)] If directly inherited, expected > dama
or rama
*daŋo ‘stinkbug’
[PWMP dbl *Zaŋaw ‘paddy-bug’ (Blust 2002)]
X
dapoʔ ‘kitchen, hearth’ cf *rapu
[< undetermined source, possibly Rkg < PMP *dapuR ‘hearth’ (Blust ACD note)] cf PSS
*(d)apuR, Mal dapur, Sad, PUS, Mmj dapo' (Mills 676)
*datu: ‘king’
[PMP *datuʔ ‘chief’ (Zorc 1995)] cf PM *datuʔ ‘head of clan’ Also, Mills (677) notes, in
addition to datu, Bug, Mak term dato', with discussion.
*maN-dulu ‘work together, help’
[PWMP *duluR ‘accompany, go together with’ (Blust ACD)]
*elu ‘saliva’
[PMP *iluR (Blust ACD)] *i- > *e- unexplained
*epiʔ ‘small cockroach sp’
[PMP *ipəs ‘cockroach’ (Blust 2002)]
*ereŋ ‘ladder, stairs’
[PMP *əZan ‘notched log ladder’ (Blust ACD)] *ə > *e irregular, expected > è; *a > *e
unexplained
*haŋ-kàliʔ ‘perhaps, in case, lest’40 ?s
[PMP *baraŋ ‘perhaps, if’ (Blust ACD)] cf Mal barangkali
*harani ‘bold, daring, savage’
[PMP *baRanih ‘hero, war leader, dare’ (Blust ACD)] *R > *r irregular. Cf PKP
*warani, irregular also
*huhu ‘fish trap’
[PMP *bubu (Blust ACD)]
X
hulu: ‘shin’ ?s
uncertain source < PKP *bulu' ‘mountain’ < PMP *bulud (Blust ACD), suggests ‘ridge’.
40
haŋkàli' can alternately be expressed saŋkàli' or ŋkàli' or kàli'. If haŋ- is derived from saŋ- ‘one’, then it is not
inherited from *baraŋ. Determining the source, however, is problematic in that the semantics of kàli' are unclear.
49
Cf PMP *lulud ‘shin’ (Blust ACD note), PKP *wuku-mbulu' ‘bone of the *bulu' ’, Blt
kaiwulu, Pen bulud ‘shin’
*maN-huni (vt) ‘hide’ cf *miN-ka-buni
[PMP *buni (Blust 1999)]
*idak ‘step on, trample’
[PMP *inZak (Blust ACD)]
*ma-ila ‘wild’
[PMP *ilah ‘wild, timorous, shy’ (Blust ACD)]
*ipo: ‘kind of poison from uncertain source, but probably from the tree, Antiaris toxicaria’
[PWMP *qipuq ‘a tree, Antiaris toxicaria, with poison sap’ (Blust ACD)]
*kanau: ‘sugar palm’
[PWMP *qanahaw (Blust ACD)] irregular *q > *k, expected < ?*kanahaw. However, a
range of other languages also show k reflexes, so the innovation *k- must have occurred
at a pre-PSek level (Mead, pers. com.)
*kandaŋ ‘drum’
[PMP *gəndaŋ (Dempwolff, in Mills 701)]
*kauʔ ‘deed’41
[PIn ?*ga(hø)u(k?) ‘work, do’ (Mills 1981 item 16)] *-k? > *-ʔ cannot be reconciled;
expect different consonant. Cf PMP *karus, kaRus, garus, gaut, kaRud ‘scratch, scrape’
(Blust ACD). PSek *kauʔ is a regular fit from any of the cited PMP etyma, but the
semantic connection is doubtful.
*saN-kìdoʔ ‘a little’
[PWMP *kədit ‘small in size or amount’ (Blust ACD)] *i > *o irregular
*maN-koko ‘reach inside and take, as rice from a cooking pot; reach inside and fondle’
[PMP (inferred) ?*kuhkuh ‘scrape’ < PAn *k-ar-uSkuS (Blust ACD)]
*maN-kutik ‘ask’
cf PSS *ko(tc)i(C) ‘press out, pick out’. Mills acknowledged possibility of PSS *kutik,
but did not reconstruct because of doubts about semantic relationship.
*lame ‘potato’
may also reference a wild tuberous plant, not well-known. Cf PSS *lame
*limmati: ‘land leech’
[PAn *Nimatək ‘leech’ (Dahl 1976), PWMP *qali-mətaq ‘paddy leech’ (Blust ACD)]
irregularities from either etymon. Expected < ?*li-matiʔ, possibly from intermediate
node, possibly indirectly inherited [check spelling whether one /m/]
*lise: ‘nit, louse egg’
[PMP *lisəhaq (Blust ACD)] presumably *lisəhaq > *lisəq > *lisiq > *lise:. PMP
doublet *ləsəq rejected because PSek reflex has no quickness on *i.
41
Derived forms of kau' ‘deed’ reference a range of meanings, including ‘do, a feast, be naughty’
50
*lite: ‘sticky sap’
[PMP *litəq ‘sap of tree or plant’ (Blust ACD)]
*loba: ‘multicolored or striped (of dog or cat)’
PKP *loba ‘spotted, dappled (of animals)’, but possibly borrowed
*lose ‘lazy’
Suggests < ?*ləRsay; analogous to PMP *bəRsay > PSek *pi-bose. Cf PKP ‘lazy’, Sad
‘weak (of one’s body)’
*mai ‘(come) hither’
[PMP *maRi ‘toward the speaker’ (Dempwolff 1938), *maRi ‘come’ (Blust 1999)]
*manaʔ ‘inheritance’
[PMP *manaq (Blust ACD), but see his note] Mills proposed PSS *mana(R?). Both these
forms irregular for PSek; expected < ?*manaC, where C is a consonant, but not *q or *R.
*mandaŋ or ‘left, remain (adv), sisa’
Possibly Xmannaŋ if inherited through PSS
*mata: ‘unripe, uncooked’
[PMP *mataq ‘green’ (Blust 1999)]
*mea: ‘strangler fig’ ?s cf *ma-lea:
[PMP *ma-qiRaq ‘red’ (Blust 1999)]42
*mi-naho: ‘slip and fall while walking’
[PMP *Nabuq ‘drop, fall’ (Dahl 1976:73-75 [check]), *nabuq ‘fall’ (Blust ACD)]
*ma-naraŋ ‘smart, clever, skilled’
[PAn *ma-Najam ‘accustomed to, tame’ (Blust ACD)]
*mi-ŋaŋa ‘agape’
[PMP *ŋaŋa (Dempwolff 1938)]
*ma-ŋura ‘young’
[PMP dbl *ŋuda (Blust ACD)]
*ohoŋ ‘shoot of palm or rattan’
[PMP *rəbuŋ ‘bamboo shoot’ (Blust pers. com. to Mead)] similar also to etymon *qubud
‘palm heart’, but *d > *ŋ makes that correspondence problematic. Note that *rə- did not
become *ò- as expected. (Try checking out with data source.)
*maN-oiʔ ‘chew sugarcane’
[PWMP *maN-ququs (Blust ACD)] vowels irregular; would fit better if < ?*maN-quqəs
*padaŋ ‘grassland’
[PAn *pa(n)daŋ (Mills)] *n > *∅
X
pasaʔ ‘market’
[< Mal pasar < (presumably) Persian bāzār]
42
Mead reconstructed *meaq for ‘red’ in PBT. That form, mea: in Seko Padang, is the generic term for ‘strangler
fig’, of which there is a red species. The semantic connection remains questionable, but consider Ram mo-mea' ‘red’
and Mmj ma-mea ‘red’.
51
X
pasiʔ ‘granulated’
[< Mal pasir ‘sand’ < PMP *pasiR ‘sand, grain’ (Dempwolff, in Zorc 1971 item 2657)]
Known only from the expression kolla pasi' ‘granulated sugar’
*pìse: ‘broken in pieces, shattered’
[PMP *pəsəq ‘break, broken’ (Zorc 1995)]
*pìsiʔ ‘stingy’
suggests < ?*pəs(i/ə)C. Cf PSS *pisA(C) ‘touch, feel (gently)’ (Mills 804), Sad pesse'
‘pinch, squeeze’. Related to PMP *pəspəs ‘squeeze’ (Zorc 1995, Blust ACD)? Indirectly
inherited from PMP *piZət (Blust ACD)?
*ma-ràba: ‘broken, damaged, rusak’ *maN-ràba: ‘dismantle’
[PMP *Rəbaq ‘destroy’ (Zorc 1971), PWMP *rəbaq ‘collapse, fall down (as a house)’
(Blust ACD)] *R/r > *r irregular; expected < PMP ?*dəbaq43
*rai ‘fine dirt on skin’
[PAn? *dakiH ‘dirt(y); body dirt, dandruff’ (Zorc 1995), PMP *daki (Blust pers. com. to
Mead)]
*rapi: ‘twin’ ?p
expected < ?*dapiq. Likely indirectly inherited. Cf PKP *rapi' ‘twins, pair (e.g., two
bananas in one skin)’, Karo Batak galuh rapit ‘double banana’ (ACD note). Probably
related to [PWMP *apij ‘twins, double banana’ (Blust ACD)] Martens (pers com)
suggests PMP *apij be revised to *-apid
*rapu ‘clan’
SekPL also, ‘hearth’, SekPH archaic, ‘hearth’ [PMP *dapuR ‘hearth’ (Blust ACD note)]
X
rarame ‘rice straw’
[probably < Sad/Rkg dadame < PMP *ZaRámi (Zorc 1995)] If directly inherited,
irregular *Z- > *r-, *R > *r, and *i > *e; expected > **da(a)mi; cf Uma, Bad deami
(Martens 1990). Cf PSS *dame (Mills 675)
*mi-raruŋ ‘feel cold’
[PMP *ma-dajəm ‘have chills’ (Blust ACD)]
*maN-rauʔ ‘grab’
[PMP *dakut (Blust ACD)]
*ma-rùbuk ‘brittle, friable, crumble’ ?p
[PMP *rəpək (Blust ACD)] *r > *r irregular; *p > *b similar, but cannot be reconciled.
Either not cognate or indirectly related. Should not be reconstructed without outside
witness
*saha ‘python’
[PMP *sawa (Blust ACD note)]
*maN-sahuʔ ‘sow, scatter, broadcast’
[PWMP *sabuD ‘sow, strew, scatter’ (Blust ACD)]
43
If PMP *r holds for *rəbaq, and *ma-ràba: is not a loan in PSek, it would stand as the only known, clear PSek *r
reflex of PMP *r.
52
*sala: ‘different, in error, mistake’
[PMP *saláq (Zorc 1995)]
*miN-sàla: ‘turn aside, swerve, deviate’
[PMP *saláq (Zorc 1995)] *a > *à irregular, expected < ?*səlaq
*maN-sàŋe ‘fry without oil’
[PMP *saŋəlaR ‘fry’ (Blust ACD)] *-laR > *∅. Else, possible loan from Mal sangra-i
‘fry without oil’ > sangai > sange
*saŋeʔ ‘forest mite?’
[PMP *səŋət ‘biting insect, wasp’ (Blust 2002)] *ə > *a irregular, expected > *à
*mi(N)-siŋka: ‘stop by, visit’
[PAn ?*si(ŋ)gaq ‘land, come ashore; visit’ (Mills 893)]
*siriŋ ‘small, biting red ant’
[PMP *si(n)jəm ‘ant sp’ (Blust ACD)] Blust (2002) reconstructed PMP *səjəm [check
with speaker to be sure no quick i]
*siso: ‘snail’
[PWMP *sisuq ‘edible snail’ (Blust ACD)]
*maN-sosok ‘insert, circulate’ ?s
[PMP *suksuk ‘stab’ (Blust 1999), dsj ‘skewer’ (Blust ACD)]
*sùdu ‘hiccup’
[Pan *sə(n)du (Blust 1986 item 292)] *n > *∅
*sukiʔ ‘rich’
[PIn ?*sugiR (Mills 835)] *R > *ʔ irregular; expected < ?*sugiC. Either *-R was
reconstructed in error, or SekP inherited this form indirectly
X
sulu: ‘torch’
< PKP language *suluq < PMP *suluq (Dempwolff, in Zorc 1971 item 3222)
*maN-suo: ‘command, order’
[PMP *suRuq (Zorc 1995)]
*ma-tahaʔ ‘tasteless, insipid’
[PMP dbl *tabaR (Blust ACD)]
*tahuni ‘afterbirth, placenta’
[PMP *tabuni (Dyen & McFarland 1970 item 658)]
*maN-tàmiʔ ‘taste’
[PMP *tamis ‘taste, try’ (Blust ACD)] *a > *à irregular, expected < ?*təmis; cf Tae
*tammi'. Possibly related to PSek *màmi' ‘sweet’
*tasak ‘done cooking’
[PMP *tasak ‘ripe’ (Charles 1973)] cf PSS *tasak ‘ripe’ (Mills 851)
*maN-tatak ‘chop up for cooking’
[PMP *təktək ‘cut (wood)’ (Blust 1999)]
53
*tìtiʔ (n) ‘drop, tetes’
[PMP dbl *titis ‘drip, leak’ (Blust ACD)] *i > *ì irregular; suggests < ?*tətis, ?*təstəs or
?*tistis. Considering Mal tetes, reconstructing PMP ?*tistis might be preferable
*maN-tòbok ‘poke’ cf Xtobok
[PMP *təbək ‘pierce, stab’ (Blust ACD)]
*tuha ‘Derris elliptica extract, fish poison’
[PMP *tuba (Zorc 1995)]
*maN-tùtuk ‘pound into pieces’
[PMP *tuktuk ‘knock, pound, beat’ (Blust 1999, dbl ACD)]
*uase ‘axe’
[PWMP *wasay (Blust ACD)]
*ubaŋ ‘gray hair’
[PMP *quban (Blust ACD)]
*uhati ‘sago grub’
[PMP dbl *qabatiR (Blust ACD); PIn ?*watɨR (Mills)] *b- > *w- > *uh-. Note similarity
to PMP *qali-wati ‘earthworm’ (Blust 1999)
54
References
Adelaar, K. Alexander. 1992. Proto Malayic: The reconstruction of its phonology and parts of its
lexicon and morphology. Pacific Linguistics C-119. Canberra: ANU.
Adelaar, K. Alexander. 1994. The classification of the Tamanic languages. In Language contact
and change in the Austronesian world, ed. by Tom Dutton and Darrell T. Tryon, 1-42.
Trends in Linguistics, Studies and Monographs 77. Berlin: Mouton de Gruyter.
Blust, Robert. 1973. Additions to ‘Proto-Austronesian addenda’ and ‘Proto-Oceanic addenda
with cognates in non-Oceanic languages’, no. II. Working Papers in Linguistics, 5.3:33-61.
Honolulu: University of Hawaii.
Blust, Robert. 1979. Proto-Western-Malayo-Polynesian Vocatives. Bijdragen tot de Taal-, Landen Volkenkunde 135:205-251.
Blust, Robert. 1986. Austronesian etymologies—III. Oceanic Linguistics 25:1-123.
Blust, Robert. 1995, in progress. Austronesian comparative dictionary. Computer files,
Department of Linguistics, University of Hawai'i.
Blust, Robert. 1999. subgrouping, circularity and extinction: some issues in Austronesian
comparative linguistics. In Selected papers from the Eighth International Conference on
Austronesian Linguistics, ed. by Elizabeth Zeitoun and Paul Jen-kuei Li, 31-94.
Symposium Series of the Institute of Linguistics, 1. Taipei: Academia Sinica.
Blust, Robert. 2001. Historical morphology and the spirit world: the *qali/kali- prefixes in
Austronesian languages. In Issues in Austronesian morphology, ed. by Joel Bradshaw and
Kenneth L. Rehg, 15-73. Pacific Linguistics 519. Canberra: ANU.
Blust, Robert. 2002. The history of faunal terms in Austronesian languages. Oceanic Linguistics
41:89-139.
Charles, Matthew. 1973. Additions not yet in list or English file. Unpublished typescript, 25pp.
Ithaca: Cornell University.
Dahl, Otto Christian. 1976. Proto-Austronesian, 2nd revised edition. Scandinavian Institute of
Asian Studies Monograph Series, 15. London: Curzon Press.
Dahl, Otto Christian. 1981a. Austronesian numerals. In Historical linguistics in Indonesia, part I,
ed. by Robert Blust, 46-58. NUSA, Linguistic Studies in Indonesian and Languages in
Indonesia 10. Jakarta: Universitas Katolik Indonesia Atma Jaya.
Dahl, Otto Christian. 1981b. Early phonetic and phonemic changes in Austronesian. Oslo:
Universitetsforlaget.
Dempwolff, Otto. 1934–1938. Vergleichende Lautlehre des uraustronesischen Wortschatzes, 3
volumes. Zeitschrift für Eingeborenen Sprachen, Beihefte 15, 17, 19. Berlin: Dietrich
Reimer.
Dyen, Isidore. 1953. The Proto-Malayo-Polynesian laryngeals. Baltimore: Linguistic Society of
America.
Dyen, Isidore and Curtis McFarland. 1970. Proto-Austronesian etyma constituting an
Austronesian cognate finder list. Unpublished typescript, 31 pp. Yale University.
Laskowske, Thomas. 1993. Fonologi bahasa Seko Padang dan penerapan teori fonologi
autosegmental. In Penyelidikan bahasa dan perkembangan wawasannya I, 256-274.
Jakarta: Masyarakat Linguistik Indonesia.
Martens, Michael. 1989. The Badaic languages. In Studies in Sulawesi linguistics, part I, ed. by
James N. Sneddon, 19-53. NUSA, Linguistic Studies of Indonesian and Other Languages
in Indonesia 31. Jakarta: Universitas Katolik Indonesia Atma Jaya.
55
Martens, Michael. 1990. Proto Kaili-Pamona: reconstruction of the protolanguage of a language
subgroup in Sulawesi. Unpublished typescript.
Mead, David. 2003. Evidence for a Celebic supergroup. In Issues in Austronesian historical
phonology, ed. by John Lynch, 115-141. Pacific Linguistics 550. Canberra: ANU.
Mills, Roger F. 1975. Proto South Sulawesi and proto Austronesian phonology. Ph.D.
dissertation, University of Michigan.
Mills, Roger F. 1981. Additional addenda. In Historical linguistics in Indonesia, part I, ed. by
Robert Blust, 59-82. NUSA, Linguistics Studies in Indonesian and Languages in Indonesia
10. Jakarta: Universitas Katolik Indonesia Atma Jaya.
Sirk, Ülo. 1989. On the evidential basis for the South Sulawesi language group. In Studies in
Sulawesi linguistics, part I, ed. by James N. Sneddon, 55-82. NUSA, Linguistic Studies of
Indonesian and Other Languages in Indonesia 31. Jakarta: Universitas Katolik Indonesia
Atma Jaya.
Smith, Beverly Kean. 1993. A phonological reconstruction of Proto-Pitu Ulunna Salu. M.A.
thesis, University of Texas at Arlington.
van den Berg. 1991. Muna dialects and Munic languages: towards a reconstruction. In Papers
from the Fifth International Conference on Austronesian Linguistics, Vol. 2 (VICAL2):
Western Austronesian and contact languages, ed. by Ray Harlow, 21-51. Aukland:
Linguistic Society of New Zealand.
Zorc, R. David. 1971. Proto-Philippine finder list. Cornell University. Unpublished typescript.
Zorc, R. David. 1995. A glossary of Austronesian reconstructions. In Comparative Austronesian
dictionary: an introduction to Austronesian studies, part 1: fascicle 2, ed. by Darrel T.
Tryon, 1105-1197. Trends in Linguistics, Documentation 10. Berlin: Mouton de Gruyter.
The preceding document was presented at the Tenth International Conference on Austronesian Linguistics (10­ICAL). To properly reference this work, please use the following format: <LastName>, <FirstName>. 2006. <PaperTitle>. Paper presented at Tenth International Conference on Austronesian Linguistics. 17­20 January 2006. Puerto Princesa City, Palawan, Philippines. http://www.sil.org/asia/philippines/ical/papers.html For other papers that were presented at 10­ICAL, please visit http://www.sil.org/asia/philippines/ical/papers.html.
Download